Telecom Churn Case Study¶

Analysis Approach :¶

Telecommunications industry experiences an average of 15 - 25% annual churn rate. Given the fact that it costs 5-10 times more to acquire a new customer than to retain an existing one, customer retention has become even more important than customer acquisition.
Here we are given with 4 months of data related to customer usage. In this case study, we analyse customer-level data of a leading telecom firm, build predictive models to identify customers at high risk of churn and identify the main indicators of churn.
Churn is predicted using two approaches. Usage based churn and Revenue based churn. Usage based churn:
Customers who have zero usage, either incoming or outgoing - in terms of calls, internet etc. over a period of time.
This case study only considers usage based churn.
In the Indian and the southeast Asian market, approximately 80% of revenue comes from the top 20% customers (called high-value customers). Thus, if we can reduce churn of the high-value customers, we will be able to reduce significant revenue leakage. Hence, this case study focuses on high value customers only.
The dataset contains customer-level information for a span of four consecutive months - June, July, August and September. The months are encoded as 6, 7, 8 and 9, respectively.
The business objective is to predict the churn in the last (i.e. the ninth) month using the data (features) from the first three months.
This is a classification problem, where we need to predict whether the customers is about to churn or not. We have carried out Baseline Logistic Regression, then Logistic Regression with PCA, PCA + Random Forest, PCA + XGBoost.

Analysis Steps¶

Data Cleaning and EDA¶

We have started with importing Necessary packages and libraries.
We have loaded the dataset into a dataframe.
We have checked the number of columns, their data types, Null count and unique value_value_count to get some understanding about data and to check if the columns are under correct data-type.
Checking for duplicate records (rows) in the data. There were no duplicates.
Since 'mobile_number' is the unique identifier available, we have made it our index to retain the identity.
Have found some columns that donot follow the naming standard, we have renamed those columns to make sure all the variables follow the same naming convention.
Follwing with column renaming, we have dealt with converting the columns into their respective data types. Here, we have evaluated all the columns which are having less than or equal to 29 unique values as catrgorical columns and rest as contineous columns.
The date columns were having 'object' as their data type, we have converted to the proper datetime format.
Since, our analysis is focused on the HVC(High value customers), we have filtered for high value customers to carryout the further analysis. The metric of this filtering of HVC is such that all the customers whose 'Average_rech_amt' of months 6 and 7 greater than or equal to 70th percentile of the 'Average_rech_amt' are considered as High Value Customers.
Checked for missing values.
Dropped all the columns with missing values greater than 50%.
We have been given 4 months data. Since each months revenue and usage data is not related to other, we did month-wise drill down on missing values.
Some columns had similar range of missing values. So, we have looked at their related columns and checked if these might be imputed with zero.
We have found that 'last_date_of_the_month' had some misisng values, so this is very meaningful and we have imputed the last date based on the month.
We have found some columns with only one unique value, so it is of no use for the analysis, hence we have dropped those columns.
Once after checking all the data preparation tasks, tagged the Churn variable(which is our target variable).
After imputing, we have dropped churn phase columns (Columns belonging to month - 9).
After all the above processing, we have retained 30,011 rows and 126 columns.
Exploratory Data Analysis

The telecom company has many users with negative average revenues in both phases. These users are likely to churn.
Most customers prefer the plans of '0' category.
The customers with lesser 'aon' are more likely to Churn when compared to the Customers with higer 'aon'.
Revenue generated by the Customers who are about to churn is very unstable.
The Customers whose arpu decreases in 7th month are more likely to churn when compared to ones with increase in arpu.
The Customers with high total_og_mou in 6th month and lower total_og_mou in 7th month are more likely to churn compared to the rest.
The Customers with decrease in rate of total_ic_mou in 7th month are more likely to churn, compared to the rest.
Customers with stable usage of 2g volume throughout 6 and 7 months are less likely to churn.
Customers with fall in usage of 2g volume in 7th month are more likely to Churn.
Customers with stable usage of 3g volume throughout 6 and 7 months are less likely to churn.
Customers with fall in consumption of 3g volume in 7th month are more likely to Churn.
The customers with lower total_og_mou in 6th and 8th months are more likely to Churn compared to the ones with higher total_og_mou.
The customers with lesser total_og_mou_8 and aon are more likely to churn compared to the one with higher total_og_mou_8 and aon.
The customers with less total_ic_mou_8 are more likely to churn irrespective of aon.
The customers with total_ic_mou_8 > 2000 are very less likely to churn.

Correlation analysis has been performed.
We have created the derived variables and then removed the variables that were used to derive new ones.
Outlier treatment has been performed. We have looked at the quantiles to understand the spread of Data.
We have capped the upper outliers to 99th percentile.
We have checked categorical variables and contribution of classes in those variables. The classes with less ccontribution are grouped into 'Others'.
Dummy Variables were created.

Pre-processing Steps¶

Train-Test Split has been performed.
The data has high class-imbalance with the ratio of 0.095 (class 1 : class 0).
SMOTE technique has been used to overcome class-imbalance.
Predictor columns have been standardized to mean - 0 and standard_deviation- 1.

Modelling¶

Model 1 : Logistic Regression with RFE & Manual Elimination ( Interpretable Model )
Most important predictors of Churn , in order of importance and their coefficients are as follows :

loc_ic_t2f_mou_8 -1.2736
total_rech_num_8 -1.2033
total_rech_num_6 0.6053
monthly_3g_8_0 0.3994
monthly_2g_8_0 0.3666
std_ic_t2f_mou_8 -0.3363
std_og_t2f_mou_8 -0.2474
const -0.2336
monthly_3g_7_0 -0.2099
std_ic_t2f_mou_7 0.1532
sachet_2g_6_0 -0.1108
sachet_2g_7_0 -0.0987
sachet_2g_8_0 0.0488
sachet_3g_6_0 -0.0399

PCA: PCA : 95% of variance in the train set can be explained by first 16 principal components and 100% of variance is explained by the first 45 principal components.

Model 2 : PCA + Logistic Regression

    Train Performance :  

    Accuracy : 0.627
    Sensitivity / True Positive Rate / Recall : 0.918
    Specificity / True Negative Rate :  0.599
    Precision / Positive Predictive Value : 0.179
    F1-score : 0.3

    Test Performance :

    Accuracy : 0.086
    Sensitivity / True Positive Rate / Recall : 1.0
    Specificity / True Negative Rate :  0.0
    Precision / Positive Predictive Value : 0.086
    F1-score : 0.158

Model 3 : PCA + Random Forest Classifier

    Train Performance :  

    Accuracy : 0.882
    Sensitivity / True Positive Rate / Recall : 0.816
    Specificity / True Negative Rate :  0.888
    Precision / Positive Predictive Value : 0.408
    F1-score : 0.544

    Test Performance :

    Accuracy : 0.86
    Sensitivity / True Positive Rate / Recall : 0.80
    Specificity / True Negative Rate :  0.78
    Precision / Positive Predictive Value :0.37
    F1-score :0.51

Model 4 : PCA + XGBoost

    Train Performance :  

    Accuracy : 0.873
    Sensitivity / True Positive Rate / Recall : 0.887
    Specificity / True Negative Rate :  0.872
    Precision / Positive Predictive Value : 0.396
    F1-score : 0.548

    Test Performance :

    Accuracy : 0.086
    Sensitivity / True Positive Rate / Recall : 1.0
    Specificity / True Negative Rate :  0.0
    Precision / Positive Predictive Value : 0.086
    F1-score : 0.158

Recommendations :¶

Following are the strongest indicators of churn

Customers who churn show lower average monthly local incoming calls from fixed line in the action period by 1.27 standard deviations , compared to users who don't churn , when all other factors are held constant. This is the strongest indicator of churn. Customers who churn show lower number of recharges done in action period by 1.20 standard deviations, when all other factors are held constant. This is the second strongest indicator of churn. Further customers who churn have done 0.6 standard deviations higher recharge than non-churn customers. This factor when coupled with above factors is a good indicator of churn. Customers who churn are more likely to be users of 'monthly 2g package-0 / monthly 3g package-0' in action period (approximately 0.3 std deviations higher than other packages), when all other factors are held constant.

Based on the above indicators the recommendations to the telecom company are :

Concentrate on users with 1.27 std devations lower than average incoming calls from fixed line. They are most likely to churn. Concentrate on users who recharge less number of times ( less than 1.2 std deviations compared to avg) in the 8th month. They are second most likely to churn. Models with high sensitivity are the best for predicting churn. Use the PCA + Logistic Regression model to predict churn. It has an ROC score of 0.87, test sensitivity of 100%.

Analysis¶

Data Understanding¶

# Importing Necessary Libraries.
import numpy as np, pandas as pd, matplotlib.pyplot as plt, seaborn as sns
import warnings 
warnings.filterwarnings('ignore')

# Setting max display columns and rows.
pd.set_option('display.max_rows', 500)
pd.set_option('display.max_columns', 500)

# Reading Dataset into a DataFrame.
data=pd.read_csv('telecom_churn_data.csv')
data.head()

# Checking information about data.
print(data.info())
def metadata_matrix(data) : 
    return pd.DataFrame({
                'Datatype' : data.dtypes.astype(str), 
                'Non_Null_Count': data.count(axis = 0).astype(int), 
                'Null_Count': data.isnull().sum().astype(int), 
                'Null_Percentage': round(data.isnull().sum()/len(data) * 100 , 2), 
                'Unique_Values_Count': data.nunique().astype(int) 
                 }).sort_values(by='Null_Percentage', ascending=False)

metadata_matrix(data)

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 99999 entries, 0 to 99998
Columns: 226 entries, mobile_number to sep_vbc_3g
dtypes: float64(179), int64(35), object(12)
memory usage: 172.4+ MB
None

Data Cleaning¶

# Checking if there are any duplicate records.
data['mobile_number'].value_counts().sum()

99999

Since number of rows is same as distinct mobile numbers, there is no duplicate data

# mobile_number is a unique identifier 
# Setting mobile_number as the index 
data = data.set_index('mobile_number')

# Renaming columns 
data = data.rename({'jun_vbc_3g' : 'vbc_3g_6', 'jul_vbc_3g' : 'vbc_3g_7', 'aug_vbc_3g' : 'vbc_3g_8', 'sep_vbc_3g' : 'vbc_3g_9'}, axis=1)

#Converting columns into appropriate data types and extracting singe value columns.
# Columns with unique values < 29 are considered as categorical variables. 
# The number 30 is arrived at, by looking at the above metadata_matrix output. 

columns=data.columns
change_to_cat=[]
single_value_col=[]
for column in columns:
    unique_value_count=data[column].nunique()
    if unique_value_count==1:
        single_value_col.append(column)
    if unique_value_count<=29 and unique_value_count!=0 and data[column].dtype in ['int','float']:
        change_to_cat.append(column)
print( ' Columns to change to categorical data type : \n' ,pd.DataFrame(change_to_cat), '\n')

 Columns to change to categorical data type : 
                    0
0          circle_id
1     loc_og_t2o_mou
2     std_og_t2o_mou
3     loc_ic_t2o_mou
4   std_og_t2c_mou_6
5   std_og_t2c_mou_7
6   std_og_t2c_mou_8
7   std_og_t2c_mou_9
8   std_ic_t2o_mou_6
9   std_ic_t2o_mou_7
10  std_ic_t2o_mou_8
11  std_ic_t2o_mou_9
12   count_rech_3g_6
13   count_rech_3g_7
14   count_rech_3g_8
15   count_rech_3g_9
16  night_pck_user_6
17  night_pck_user_7
18  night_pck_user_8
19  night_pck_user_9
20      monthly_2g_6
21      monthly_2g_7
22      monthly_2g_8
23      monthly_2g_9
24      monthly_3g_6
25      monthly_3g_7
26      monthly_3g_8
27      monthly_3g_9
28       sachet_3g_6
29       sachet_3g_7
30       sachet_3g_8
31       sachet_3g_9
32         fb_user_6
33         fb_user_7
34         fb_user_8
35         fb_user_9

# Converting all the above columns having <=29 unique values into categorical data type.
data[change_to_cat]=data[change_to_cat].astype('category')

# Converting *sachet* variables to categorical data type 
sachet_columns = data.filter(regex='.*sachet.*', axis=1).columns.values
data[sachet_columns] = data[sachet_columns].astype('category')

#Changing datatype of date variables to datetime.
columns=data.columns
col_with_date=[]
import re
for column in columns:
    x = re.findall("^date", column)
    if x:
        col_with_date.append(column)
data[col_with_date].dtypes

date_of_last_rech_6         object
date_of_last_rech_7         object
date_of_last_rech_8         object
date_of_last_rech_9         object
date_of_last_rech_data_6    object
date_of_last_rech_data_7    object
date_of_last_rech_data_8    object
date_of_last_rech_data_9    object
dtype: object

# Checking the date format
data[col_with_date].head()

Lets convert the above columns to datetime data type.

for col in col_with_date:
    data[col]=pd.to_datetime(data[col], format="%m/%d/%Y")
data[col_with_date].head()

Filtering High Value Customers¶

Customers are High Values if their Average recharge amount of june and july is more than or equal to 70th percentile of Average recharge amount.

#Deriving Average recharge amount of June and July.
data['Average_rech_amt_6n7']=(data['total_rech_amt_6']+data['total_rech_amt_7'])/2

#Filtering based HIGH VALUED CUSTOMERS based on (Average_rech_amt_6n7 >= 70th percentile of Average_rech_amt_6n7)
data=data[(data['Average_rech_amt_6n7']>= data['Average_rech_amt_6n7'].quantile(0.7))]

Missing Values¶

#Checking for missing values.
missing_values = metadata_matrix(data)[['Datatype', 'Null_Percentage']].sort_values(by='Null_Percentage', ascending=False)
missing_values

# Columns with high missing values , > 50%
metadata = metadata_matrix(data)
condition = metadata['Null_Percentage'] > 50 
high_missing_values = metadata[condition]
high_missing_values

# Dropping above columns with high missing values 
high_missing_value_columns = high_missing_values.index 
data.drop(columns=high_missing_value_columns, inplace=True)

# Looking at remaining columns with missing values 
metadata_matrix(data)

data contains information of 04 months - 6,7,8,9.
For the purpose of missing value treatment, each month's revenue and usage data is not related to the other months.
hence, missing value treatment could be performed month wise.

# Month 6

sixth_month_columns = []
for column in data.columns:
    x = re.search("6$", column)
    if x:
        sixth_month_columns.append(column)
# missing_values.loc[sixth_month_columns].sort_values(by='Null_Percentage', ascending=False)
metadata = metadata_matrix(data)
condition = metadata.index.isin(sixth_month_columns)
sixth_month_metadata = metadata[condition]
sixth_month_metadata

Note that all the columns with *_mou have exactly 3.94% rows with missing values.
This is an indicator of a meaningful missing values.
Further note that *_mou columns indicate minutes of usage, which are applicable only to customers using calling plans. It is probable that, the 3.94% customers not using calling plans.
This could confirmed by looking at 'total_og_mou_6' and 'total_ic_mou_6' related columns where _mou columns have missing values. If these columns are zero for a customer , then all _mou columns should be zero too.

#  columns with meaningful missing in 6th month 
sixth_month_meaningful_missing_condition = sixth_month_metadata['Null_Percentage'] == 1.05
sixth_month_meaningful_missing_cols = sixth_month_metadata[sixth_month_meaningful_missing_condition].index.values
sixth_month_meaningful_missing_cols

array(['std_ic_mou_6', 'offnet_mou_6', 'std_ic_t2f_mou_6', 'isd_ic_mou_6',
       'ic_others_6', 'onnet_mou_6', 'std_ic_t2m_mou_6',
       'loc_ic_t2t_mou_6', 'loc_ic_t2m_mou_6', 'loc_ic_t2f_mou_6',
       'loc_ic_mou_6', 'std_ic_t2t_mou_6', 'og_others_6', 'spl_og_mou_6',
       'roam_ic_mou_6', 'spl_ic_mou_6', 'std_og_t2t_mou_6',
       'loc_og_t2c_mou_6', 'std_og_t2m_mou_6', 'loc_og_t2f_mou_6',
       'std_og_t2f_mou_6', 'loc_og_t2m_mou_6', 'std_ic_t2o_mou_6',
       'std_og_t2c_mou_6', 'std_og_mou_6', 'loc_og_t2t_mou_6',
       'isd_og_mou_6', 'roam_og_mou_6', 'loc_og_mou_6'], dtype=object)

# Looking at all sixth month columns where rows of *_mou are null
condition = data[sixth_month_meaningful_missing_cols].isnull()
# data.loc[condition, sixth_month_columns]


# Rows is null for all the above columns 
missing_rows = pd.Series([True]*data.shape[0], index = data.index)
for column in sixth_month_meaningful_missing_cols : 
    missing_rows = missing_rows & data[column].isnull()

print('Total outgoing mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_og_mou_6'].unique()[0])
print('Total incoming mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_ic_mou_6'].unique()[0])

Total outgoing mou for each customer with missing *_mou data is  0.0
Total incoming mou for each customer with missing *_mou data is  0.0

Hence, these could be imputed with 0

# Imputation
data[sixth_month_meaningful_missing_cols] = data[sixth_month_meaningful_missing_cols].fillna(0)

metadata = metadata_matrix(data)

# Remaining Missing Values
metadata.iloc[metadata.index.isin(sixth_month_columns)]

Looks like there '1.61%' customers with missing date of last recharge. Let's look at 'recharge' related columns for such customers

# Looking at 'recharge' related 6th month columns for customers with missing 'date_of_last_rech_6' 
condition = data['date_of_last_rech_6'].isnull()
data[condition].filter(regex='.*rech.*6$', axis=1).head()

data[condition].filter(regex='.*rech.*6$', axis=1).nunique()

total_rech_num_6       1
total_rech_amt_6       1
max_rech_amt_6         1
date_of_last_rech_6    0
dtype: int64

Notice, that the recharge related columns for customers with missing 'date_of_last_rech_6' has just one unique value. From the first few rows of the output, we see that this is 0.
Hence, 'date_of_last_rech_6' is missing since there were no recharges made in this month.
These are meaning missing values

# Check for missing values in 6th month variables
metadata = metadata_matrix(data)
metadata[metadata.index.isin(sixth_month_columns)]

No more Missing Values in 6th month columns

# Month : 7 
seventh_month_columns = data.filter(regex='7$', axis=1).columns
seventh_month_columns

Index(['last_date_of_month_7', 'arpu_7', 'onnet_mou_7', 'offnet_mou_7',
       'roam_ic_mou_7', 'roam_og_mou_7', 'loc_og_t2t_mou_7',
       'loc_og_t2m_mou_7', 'loc_og_t2f_mou_7', 'loc_og_t2c_mou_7',
       'loc_og_mou_7', 'std_og_t2t_mou_7', 'std_og_t2m_mou_7',
       'std_og_t2f_mou_7', 'std_og_t2c_mou_7', 'std_og_mou_7', 'isd_og_mou_7',
       'spl_og_mou_7', 'og_others_7', 'total_og_mou_7', 'loc_ic_t2t_mou_7',
       'loc_ic_t2m_mou_7', 'loc_ic_t2f_mou_7', 'loc_ic_mou_7',
       'std_ic_t2t_mou_7', 'std_ic_t2m_mou_7', 'std_ic_t2f_mou_7',
       'std_ic_t2o_mou_7', 'std_ic_mou_7', 'total_ic_mou_7', 'spl_ic_mou_7',
       'isd_ic_mou_7', 'ic_others_7', 'total_rech_num_7', 'total_rech_amt_7',
       'max_rech_amt_7', 'date_of_last_rech_7', 'last_day_rch_amt_7',
       'vol_2g_mb_7', 'vol_3g_mb_7', 'monthly_2g_7', 'sachet_2g_7',
       'monthly_3g_7', 'sachet_3g_7', 'vbc_3g_7', 'Average_rech_amt_6n7'],
      dtype='object')

seventh_month_metadata = metadata[metadata.index.isin(seventh_month_columns)]
seventh_month_metadata

Note that all the columns with *_mou have exactly 3.86% rows with missing values.
This is an indicator of a meaningful missing values.
Further note that *_mou columns indicate minutes of usage, which are applicable only to customers using calling plans. It is probable that, the 3.86% customers not using calling plans.
This could confirmed by looking at 'total_og_mou_7' and 'total_ic_mou_7' related columns where _mou columns have missing values. If these columns are zero for a customer , then all _mou columns should be zero too.

#  columns with meaningful missing in 7th month 
seventh_month_meaningful_missing_condition = seventh_month_metadata['Null_Percentage'] == 1.01
seventh_month_meaningful_missing_cols = seventh_month_metadata[seventh_month_meaningful_missing_condition].index.values
seventh_month_meaningful_missing_cols

array(['loc_ic_t2t_mou_7', 'og_others_7', 'loc_ic_t2f_mou_7',
       'loc_ic_t2m_mou_7', 'loc_ic_mou_7', 'std_ic_t2t_mou_7',
       'std_ic_t2f_mou_7', 'std_ic_t2o_mou_7', 'std_ic_mou_7',
       'spl_ic_mou_7', 'isd_ic_mou_7', 'ic_others_7', 'std_ic_t2m_mou_7',
       'isd_og_mou_7', 'spl_og_mou_7', 'std_og_t2f_mou_7', 'onnet_mou_7',
       'offnet_mou_7', 'roam_ic_mou_7', 'roam_og_mou_7',
       'loc_og_t2t_mou_7', 'loc_og_t2f_mou_7', 'loc_og_t2c_mou_7',
       'loc_og_mou_7', 'std_og_t2t_mou_7', 'std_og_t2m_mou_7',
       'loc_og_t2m_mou_7', 'std_og_t2c_mou_7', 'std_og_mou_7'],
      dtype=object)

# Looking at all 7th month columns where rows of *_mou are null
condition = data[seventh_month_meaningful_missing_cols].isnull()

# Rows is null for all the above columns 
missing_rows = pd.Series([True]*data.shape[0], index = data.index)
for column in seventh_month_meaningful_missing_cols : 
    missing_rows = missing_rows & data[column].isnull()

print('Total outgoing mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_og_mou_7'].unique()[0])
print('Total incoming mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_ic_mou_7'].unique()[0])

Total outgoing mou for each customer with missing *_mou data is  0.0
Total incoming mou for each customer with missing *_mou data is  0.0

Hence, these could be imputed with 0

# Imputation
data[seventh_month_meaningful_missing_cols] = data[seventh_month_meaningful_missing_cols].fillna(0)

metadata = metadata_matrix(data)

# Remaining Missing Values
metadata.iloc[metadata.index.isin(seventh_month_columns)]

Looks like there '1.77%' customers with missing date of last recharge. Let's look at 'recharge' related columns for such customers

# Looking at 'recharge' related 7th month columns for customers with missing 'date_of_last_rech_7' 
condition = data['date_of_last_rech_7'].isnull()
data[condition].filter(regex='.*rech.*7$', axis=1).head()

data[condition].filter(regex='.*rech.*7$', axis=1).nunique()

total_rech_num_7         1
total_rech_amt_7         1
max_rech_amt_7           1
date_of_last_rech_7      0
Average_rech_amt_6n7    90
dtype: int64

Notice, that the recharge related columns for customers with missing 'date_of_last_rech_7' has just one unique value. From the first few rows of the output, we see that this is 0.
Hence, 'date_of_last_rech_7' is missing since there were no recharges made in this month.
These are meaning missing values

# Month : 8

eighth_month_columns = data.filter(regex="8$", axis=1).columns
metadata = metadata_matrix(data)
condition = metadata.index.isin(eighth_month_columns)
eighth_month_metadata = metadata[condition]
eighth_month_metadata

#  columns with meaningful missing in 8th month 
eighth_month_meaningful_missing_condition = eighth_month_metadata['Null_Percentage'] == 3.13
eighth_month_meaningful_missing_cols = eighth_month_metadata[eighth_month_meaningful_missing_condition].index.values
eighth_month_meaningful_missing_cols

array(['std_og_t2c_mou_8', 'std_og_mou_8', 'isd_og_mou_8', 'loc_ic_mou_8',
       'std_og_t2m_mou_8', 'loc_ic_t2m_mou_8', 'loc_og_mou_8',
       'std_og_t2t_mou_8', 'std_og_t2f_mou_8', 'loc_ic_t2f_mou_8',
       'loc_og_t2c_mou_8', 'ic_others_8', 'loc_og_t2m_mou_8',
       'spl_og_mou_8', 'roam_ic_mou_8', 'std_ic_mou_8', 'spl_ic_mou_8',
       'std_ic_t2o_mou_8', 'onnet_mou_8', 'loc_og_t2f_mou_8',
       'offnet_mou_8', 'std_ic_t2f_mou_8', 'og_others_8',
       'loc_ic_t2t_mou_8', 'std_ic_t2m_mou_8', 'std_ic_t2t_mou_8',
       'roam_og_mou_8', 'isd_ic_mou_8', 'loc_og_t2t_mou_8'], dtype=object)

# Looking at all 8th month columns where rows of *_mou are null
condition = data[eighth_month_meaningful_missing_cols].isnull()

# Rows is null for all the above columns 
missing_rows = pd.Series([True]*data.shape[0], index = data.index)
for column in eighth_month_meaningful_missing_cols : 
    missing_rows = missing_rows & data[column].isnull()

print('Total outgoing mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_og_mou_8'].unique()[0])
print('Total incoming mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_ic_mou_8'].unique()[0])

Total outgoing mou for each customer with missing *_mou data is  0.0
Total incoming mou for each customer with missing *_mou data is  0.0

# Imputation
data[eighth_month_meaningful_missing_cols] = data[eighth_month_meaningful_missing_cols].fillna(0)

metadata = metadata_matrix(data)

# Remaining Missing Values
metadata.iloc[metadata.index.isin(eighth_month_columns)]

# Looking at 'recharge' related 8th month columns for customers with missing 'date_of_last_rech_8' 
condition = data['date_of_last_rech_8'].isnull()
data[condition].filter(regex='.*rech.*8$', axis=1).head()

data[condition].filter(regex='.*rech.*8$', axis=1).nunique()

total_rech_num_8       1
total_rech_amt_8       1
max_rech_amt_8         1
date_of_last_rech_8    0
dtype: int64

# Month : 9

ninth_month_columns = data.filter(regex="9$", axis=1).columns
metadata = metadata_matrix(data)
condition = metadata.index.isin(ninth_month_columns)
ninth_month_metadata = metadata[condition]
ninth_month_metadata

#  columns with meaningful missing in 9th month 
ninth_month_meaningful_missing_condition = ninth_month_metadata['Null_Percentage'] == 5.68
ninth_month_meaningful_missing_cols = ninth_month_metadata[ninth_month_meaningful_missing_condition].index.values
ninth_month_meaningful_missing_cols

array(['std_og_t2c_mou_9', 'spl_ic_mou_9', 'loc_og_t2m_mou_9',
       'og_others_9', 'loc_og_t2c_mou_9', 'isd_ic_mou_9',
       'loc_og_t2t_mou_9', 'spl_og_mou_9', 'loc_ic_t2t_mou_9',
       'loc_og_mou_9', 'roam_og_mou_9', 'std_ic_mou_9',
       'loc_ic_t2m_mou_9', 'roam_ic_mou_9', 'std_og_t2t_mou_9',
       'offnet_mou_9', 'loc_ic_t2f_mou_9', 'std_ic_t2f_mou_9',
       'isd_og_mou_9', 'std_og_mou_9', 'std_og_t2f_mou_9', 'ic_others_9',
       'std_ic_t2t_mou_9', 'std_ic_t2o_mou_9', 'loc_og_t2f_mou_9',
       'std_og_t2m_mou_9', 'loc_ic_mou_9', 'std_ic_t2m_mou_9',
       'onnet_mou_9'], dtype=object)

# Looking at all 9th month columns where rows of *_mou are null
condition = data[ninth_month_meaningful_missing_cols].isnull()

# Rows is null for all the above columns 
missing_rows = pd.Series([True]*data.shape[0], index = data.index)
for column in ninth_month_meaningful_missing_cols : 
    missing_rows = missing_rows & data[column].isnull()

print('Total outgoing mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_og_mou_9'].unique()[0])
print('Total incoming mou for each customer with missing *_mou data is ', data.loc[missing_rows,'total_ic_mou_9'].unique()[0])

Total outgoing mou for each customer with missing *_mou data is  0.0
Total incoming mou for each customer with missing *_mou data is  0.0

# Imputation
data[ninth_month_meaningful_missing_cols] = data[ninth_month_meaningful_missing_cols].fillna(0)

metadata = metadata_matrix(data)

# Remaining Missing Values
metadata.iloc[metadata.index.isin(ninth_month_columns)]

# Looking at 'recharge' related 9th month columns for customers with missing 'date_of_last_rech_9' 
condition = data['date_of_last_rech_9'].isnull()
data[condition].filter(regex='.*rech.*9$', axis=1).head()

data[condition].filter(regex='.*rech.*9$', axis=1).nunique()

total_rech_num_9       1
total_rech_amt_9       1
max_rech_amt_9         1
date_of_last_rech_9    0
dtype: int64

# Imputing "last_date_of_month_*"

print('Missing Value Percentage in last_date_of_month columns : \n', 100*data.filter(regex='last_date_of_month_.*', axis=1).isnull().sum() / data.shape[0], '\n')
print('The unique values in last_date_of_month_6 : ' , data['last_date_of_month_6'].unique())
print('The unique values in last_date_of_month_7 : ' , data['last_date_of_month_7'].unique())
print('The unique values in last_date_of_month_8 : ' , data['last_date_of_month_8'].unique())
print('The unique values in last_date_of_month_9 : ' , data['last_date_of_month_9'].unique())

Missing Value Percentage in last_date_of_month columns : 
 last_date_of_month_6    0.000000
last_date_of_month_7    0.103295
last_date_of_month_8    0.523142
last_date_of_month_9    1.199560
dtype: float64 

The unique values in last_date_of_month_6 :  ['6/30/2014']
The unique values in last_date_of_month_7 :  ['7/31/2014' nan]
The unique values in last_date_of_month_8 :  ['8/31/2014' nan]
The unique values in last_date_of_month_9 :  ['9/30/2014' nan]

Last date of month is the last calender date of a particular month, it is independent of the churn data.
Lets impute these missing values using mode.

# Imputing last_date_of_month_* values
data['last_date_of_month_7'] = data['last_date_of_month_7'].fillna(data['last_date_of_month_7'].mode()[0])
data['last_date_of_month_8'] = data['last_date_of_month_8'].fillna(data['last_date_of_month_8'].mode()[0])
data['last_date_of_month_9'] = data['last_date_of_month_9'].fillna(data['last_date_of_month_9'].mode()[0])

data['last_date_of_month_7'].unique()

array(['7/31/2014'], dtype=object)

metadata = metadata_matrix(data)
metadata

print(data[data['date_of_last_rech_6'].isnull()][['date_of_last_rech_6','total_rech_amt_6','total_rech_num_6']].nunique())
print(data[data['date_of_last_rech_7'].isnull()][['date_of_last_rech_7','total_rech_amt_7','total_rech_num_7']].nunique())
print(data[data['date_of_last_rech_8'].isnull()][['date_of_last_rech_8','total_rech_amt_8','total_rech_num_8']].nunique())
print(data[data['date_of_last_rech_9'].isnull()][['date_of_last_rech_9','total_rech_amt_9','total_rech_num_9']].nunique())

date_of_last_rech_6    0
total_rech_amt_6       1
total_rech_num_6       1
dtype: int64
date_of_last_rech_7    0
total_rech_amt_7       1
total_rech_num_7       1
dtype: int64
date_of_last_rech_8    0
total_rech_amt_8       1
total_rech_num_8       1
dtype: int64
date_of_last_rech_9    0
total_rech_amt_9       1
total_rech_num_9       1
dtype: int64

print("\n",data[data['date_of_last_rech_6'].isnull()][['total_rech_amt_6','total_rech_num_6']].head())
print("\n",data[data['date_of_last_rech_7'].isnull()][['total_rech_amt_7','total_rech_num_7']].head())
print("\n",data[data['date_of_last_rech_8'].isnull()][['total_rech_amt_8','total_rech_num_8']].head())
print("\n",data[data['date_of_last_rech_9'].isnull()][['total_rech_amt_9','total_rech_num_9']].head())

                total_rech_amt_6  total_rech_num_6
mobile_number                                    
7001588448                    0                 0
7001223277                    0                 0
7000721536                    0                 0
7001490351                    0                 0
7000665415                    0                 0

                total_rech_amt_7  total_rech_num_7
mobile_number                                    
7000369789                    0                 0
7001967148                    0                 0
7000066601                    0                 0
7001189556                    0                 0
7002024450                    0                 0

                total_rech_amt_8  total_rech_num_8
mobile_number                                    
7000340381                    0                 0
7000608224                    0                 0
7000369789                    0                 0
7000248548                    0                 0
7001967063                    0                 0

                total_rech_amt_9  total_rech_num_9
mobile_number                                    
7000340381                    0                 0
7000854899                    0                 0
7000369789                    0                 0
7001967063                    0                 0
7000066601                    0                 0

The columns 'date_of_last_rech' for june,july and August does not have any value becuase there are no recharges done by the user during those months.

Dropping columns with one unique value.¶

metadata=metadata_matrix(data)
singular_value_cols=metadata[metadata['Unique_Values_Count']==1].index.values
#data.loc[metadata_matrix(data)['Unique_Values_Count']==1].index

#Dropping singular value columns.
data.drop(columns=singular_value_cols,inplace=True)

# Dropping date columns 
# since they are not usage related columns and can't be used for modelling 
date_columns = data.filter(regex='^date.*').columns
data.drop(columns=date_columns, inplace=True)
metadata_matrix(data)

Tagging Churn (TARGET variable)¶

data['Churn'] = 0
churned_customers = data.query('total_og_mou_9 == 0 & total_ic_mou_9 == 0 & vol_2g_mb_9 == 0 &  vol_3g_mb_9 == 0').index
data.loc[churned_customers,'Churn']=1
data['Churn'] = data['Churn'].astype('category')

# Churn proportions
data['Churn'].value_counts(normalize=True).to_frame()

Dropping Churn Phase Columns¶

churn_phase_columns = data.filter(regex='9$').columns
data.drop(columns=churn_phase_columns, inplace=True)
print('Retained Columns')
data.columns.to_frame(index=False)

Retained Columns

print('retained no of rows', data.shape[0])
print('retain no of columns', data.shape[1])

retained no of rows 30011
retain no of columns 126

Exploratory Data Analysis¶

Summary Statistics¶

data.describe()

The telecom company has many users with negative average revenues in both phases. These users are likely to churn

categorical_columns = data.dtypes[data.dtypes == 'category'].index.values
print('Mode : ')
data[categorical_columns].mode().T

Mode :

Most customers prefer the plans of '0' category

Univariate Analysis¶

churned_customers = data[data['Churn'] == 1]
non_churned_customers = data[data['Churn'] == 0]

Age on Network¶

plt.figure(figsize=(12,8))
sns.violinplot(x='aon', y='Churn', data=data)
plt.title('Age on Network vs Churn')
plt.show()

The customers with lesser 'aon' are more likely to Churn when compared to the Customers with higer 'aon'

# function for numerical variable univariate analysis
from tabulate import tabulate
def num_univariate_analysis(column_names,scale='linear') : 
    # boxplot for column vs target
    
    fig = plt.figure(figsize=(16,8))
    ax1 = fig.add_subplot(1,3,1)
    sns.violinplot(x='Churn', y = column_names[0], data = data, ax=ax1)
    title = ''.join(column_names[0]) +' vs Churn'
    ax1.set(title=title)
    if scale == 'log' :
        plt.yscale('log')
        ax1.set(ylabel= column_names[0] + '(Log Scale)')
        
    ax2 = fig.add_subplot(1,3,2)
    sns.violinplot(x='Churn', y = column_names[1], data = data, ax=ax2)
    title = ''.join(column_names[1]) +' vs Churn'
    ax2.set(title=title)
    if scale == 'log' :
        plt.yscale('log')
        ax2.set(ylabel= column_names[1] + '(Log Scale)')
    
    ax3 = fig.add_subplot(1,3,3)
    sns.violinplot(x='Churn', y = column_names[2], data = data, ax=ax3)
    title = ''.join(column_names[2]) +' vs Churn'
    ax3.set(title=title)
    if scale == 'log' :
        plt.yscale('log')
        ax3.set(ylabel= column_names[2] + '(Log Scale)')
    
    # summary statistic
    
    print('Customers who churned (Churn : 1)')
    print(churned_customers[column_names].describe())
    
    print('\nCustomers who did not churn (Churn : 0)')
    print(non_churned_customers[column_names].describe(),'\n')

# function for categorical variable univariate analysis
!pip install sidetable
import sidetable
def cat_univariate_analysis(column_names,figsize=(16,4)) : 
    
    # column vs target count plot
    fig = plt.figure(figsize=figsize)
    
    ax1 = fig.add_subplot(1,3,1)
    sns.countplot(x=column_names[0],hue='Churn',data=data, ax=ax1)
    title = column_names[0] + ' vs No of Churned Customers'
    ax1.set(title= title)
    ax1.legend(loc='upper right')

    
    ax2 = fig.add_subplot(1,3,2)
    sns.countplot(x=column_names[1],hue='Churn',data=data, ax=ax2)
    title = column_names[1] + ' vs No of Churned Customers'
    ax2.set(title= title)
    ax2.legend(loc='upper right')

        
    ax3 = fig.add_subplot(1,3,3)
    sns.countplot(x=column_names[2],hue='Churn',data=data, ax=ax3)
    title = column_names[2] + ' vs No of Churned Customers'
    ax3.set(title= title)
    ax3.legend(loc='upper right')

        
    # Percentages 
    print('Customers who churned (Churn : 1)')
    print(tabulate(pd.DataFrame(churned_customers.stb.freq([column_names[0]])), headers='keys', tablefmt='psql'),'\n')
    print(tabulate(pd.DataFrame(churned_customers.stb.freq([column_names[1]])), headers='keys', tablefmt='psql'),'\n')
    print(tabulate(pd.DataFrame(churned_customers.stb.freq([column_names[2]])), headers='keys', tablefmt='psql'),'\n')

    print('\nCustomers who did not churn (Churn : 0)')
    print(tabulate(pd.DataFrame(non_churned_customers.stb.freq([column_names[0]])), headers='keys', tablefmt='psql'),'\n')
    print(tabulate(pd.DataFrame(non_churned_customers.stb.freq([column_names[1]])), headers='keys', tablefmt='psql'),'\n')
    print(tabulate(pd.DataFrame(non_churned_customers.stb.freq([column_names[2]])), headers='keys', tablefmt='psql'),'\n')

Requirement already satisfied: sidetable in /Users/UMAER/Documents/DataScience/anaconda3/lib/python3.8/site-packages (0.7.0)
Requirement already satisfied: pandas>=1.0 in /Users/UMAER/Documents/DataScience/anaconda3/lib/python3.8/site-packages (from sidetable) (1.0.5)
Requirement already satisfied: pytz>=2017.2 in /Users/UMAER/Documents/DataScience/anaconda3/lib/python3.8/site-packages (from pandas>=1.0->sidetable) (2020.1)
Requirement already satisfied: numpy>=1.13.3 in /Users/UMAER/Documents/DataScience/anaconda3/lib/python3.8/site-packages (from pandas>=1.0->sidetable) (1.18.5)
Requirement already satisfied: python-dateutil>=2.6.1 in /Users/UMAER/Documents/DataScience/anaconda3/lib/python3.8/site-packages (from pandas>=1.0->sidetable) (2.8.1)
Requirement already satisfied: six>=1.5 in /Users/UMAER/Documents/DataScience/anaconda3/lib/python3.8/site-packages (from python-dateutil>=2.6.1->pandas>=1.0->sidetable) (1.15.0)

arpu_6, arpu_7 , arpu_8¶

columns = ['arpu_6','arpu_7','arpu_8']
num_univariate_analysis(columns,'log')

Customers who churned (Churn : 1)
             arpu_6        arpu_7       arpu_8
count   2593.000000   2593.000000  2593.000000
mean     678.716970    550.511946   243.063343
std      551.792864    517.241221   378.843531
min     -209.465000   -158.963000   -37.887000
25%      396.507000    289.641000     0.000000
50%      573.396000    464.674000   101.894000
75%      819.460000    691.588000   351.028000
max    11505.508000  13224.119000  5228.826000

Customers who did not churn (Churn : 0)
             arpu_6        arpu_7        arpu_8
count  27418.000000  27418.000000  27418.000000
mean     578.637360    592.788162    562.453248
std      429.988265    457.265996    492.802655
min    -2258.709000  -2014.045000   -945.808000
25%      362.218000    369.610500    319.118500
50%      489.324000    496.182500    471.024000
75%      690.891750    701.418000    690.921000
max    27731.088000  35145.834000  33543.624000

We can understand from the above plots that revenue generated by the Customers who are about to churn is very unstable.
The Customers whose arpu decreases in 7th month are more likely to churn when compared to ones with increase in arpu.

total_og_mou_6, total_og_mou_7, total_og_mou_8¶

columns = ['total_og_mou_6', 'total_og_mou_7', 'total_og_mou_8']
num_univariate_analysis(columns)

Customers who churned (Churn : 1)
       total_og_mou_6  total_og_mou_7  total_og_mou_8
count     2593.000000     2593.000000     2593.000000
mean       867.961342      677.868909      225.083741
std        852.697688      786.961399      471.672718
min          0.000000        0.000000        0.000000
25%        277.880000      110.090000        0.000000
50%        658.360000      466.910000        0.000000
75%       1209.040000      926.760000      255.810000
max       8488.360000     8285.640000     5206.210000

Customers who did not churn (Churn : 0)
       total_og_mou_6  total_og_mou_7  total_og_mou_8
count    27418.000000    27418.000000    27418.000000
mean       669.554896      712.080684      661.480046
std        636.531612      674.580516      691.079113
min          0.000000        0.000000        0.000000
25%        265.682500      284.500000      227.970000
50%        500.410000      529.935000      470.475000
75%        872.070000      931.197500      866.045000
max      10674.030000    11365.310000    14043.060000

The Customers with high total_og_mou in 6th month and lower total_og_mou in 7th month are more likely to churn compared to the rest.

'total_ic_mou_6', 'total_ic_mou_7', 'total_ic_mou_8'¶

columns = ['total_ic_mou_6', 'total_ic_mou_7', 'total_ic_mou_8']
num_univariate_analysis(columns)

Customers who churned (Churn : 1)
       total_ic_mou_6  total_ic_mou_7  total_ic_mou_8
count     2593.000000     2593.000000     2593.000000
mean       241.954404      193.341076       68.807042
std        360.836586      318.183813      154.450340
min          0.000000        0.000000        0.000000
25%         49.460000       27.890000        0.000000
50%        137.330000       99.980000        0.000000
75%        289.510000      235.740000       70.290000
max       6633.180000     5137.560000     1859.280000

Customers who did not churn (Churn : 0)
       total_ic_mou_6  total_ic_mou_7  total_ic_mou_8
count    27418.000000    27418.000000    27418.000000
mean       313.712052      326.369333      316.858595
std        360.580253      372.112086      366.818717
min          0.000000        0.000000        0.000000
25%         94.460000      107.802500       98.265000
50%        212.160000      222.290000      212.360000
75%        401.602500      410.182500      402.270000
max       6798.640000     7279.080000     5990.710000

The Customers with decrease in rate of total_ic_mou in 7th month are more likely to churn, compared to the rest.

vol_2g_mb_6, vol_2g_mb_7, vol_2g_mb_8¶

columns = ['vol_2g_mb_6', 'vol_2g_mb_7', 'vol_2g_mb_8']
num_univariate_analysis(columns, 'log')

Customers who churned (Churn : 1)
       vol_2g_mb_6  vol_2g_mb_7  vol_2g_mb_8
count  2593.000000  2593.000000  2593.000000
mean     60.775588    49.054393    15.283185
std     243.084276   219.485813   120.975111
min       0.000000     0.000000     0.000000
25%       0.000000     0.000000     0.000000
50%       0.000000     0.000000     0.000000
75%       0.000000     0.000000     0.000000
max    4017.160000  3430.730000  3349.190000

Customers who did not churn (Churn : 0)
        vol_2g_mb_6   vol_2g_mb_7   vol_2g_mb_8
count  27418.000000  27418.000000  27418.000000
mean      80.569210     80.925060     74.309036
std      280.420463    285.265125    277.889339
min        0.000000      0.000000      0.000000
25%        0.000000      0.000000      0.000000
50%        0.000000      0.000000      0.000000
75%       16.937500     18.267500     14.245000
max    10285.900000   7873.550000  11117.610000

Customers with stable usage of 2g volumes throughout 6 and 7 months are less likely to churn.
Customers with fall in consumption of 2g volumes in 7th month are more likely to Churn.

vol_3g_mb_6, vol_3g_mb_7, vol_3g_mb_8, monthly_3g_6¶

columns = ['vol_3g_mb_6', 'vol_3g_mb_7', 'vol_3g_mb_8', 'monthly_3g_6']
num_univariate_analysis(columns, 'log')

Customers who churned (Churn : 1)
       vol_3g_mb_6   vol_3g_mb_7   vol_3g_mb_8
count  2593.000000   2593.000000   2593.000000
mean    188.395461    157.714254     56.776880
std     715.327843    690.773561    446.532769
min       0.000000      0.000000      0.000000
25%       0.000000      0.000000      0.000000
50%       0.000000      0.000000      0.000000
75%       0.000000      0.000000      0.000000
max    9400.120000  15115.510000  13440.720000

Customers who did not churn (Churn : 0)
        vol_3g_mb_6   vol_3g_mb_7   vol_3g_mb_8
count  27418.000000  27418.000000  27418.000000
mean     265.012522    289.478375    290.016390
std      878.846885    868.808831    885.821105
min        0.000000      0.000000      0.000000
25%        0.000000      0.000000      0.000000
50%        0.000000      0.000000      0.000000
75%        0.000000     35.855000     27.120000
max    45735.400000  28144.120000  30036.060000

Customers with stable usage of 3g volumes throughout 6 and 7 months are less likely to churn.
Customers with fall in consumption of 3g volumes in 7th month are more likely to Churn.

monthly_2g_6, monthly_2g_7, monthly_2g_8¶

columns = ['monthly_2g_6', 'monthly_2g_7', 'monthly_2g_8']
cat_univariate_analysis(columns)

Customers who churned (Churn : 1)
+----+----------------+---------+------------+--------------------+----------------------+
|    |   monthly_2g_6 |   count |    percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+------------+--------------------+----------------------|
|  0 |              0 |    2454 | 94.6394    |               2454 |              94.6394 |
|  1 |              1 |     126 |  4.85924   |               2580 |              99.4987 |
|  2 |              2 |      11 |  0.424219  |               2591 |              99.9229 |
|  3 |              4 |       2 |  0.0771307 |               2593 |             100      |
+----+----------------+---------+------------+--------------------+----------------------+ 

+----+----------------+---------+-----------+--------------------+----------------------+
|    |   monthly_2g_7 |   count |   percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+-----------+--------------------+----------------------|
|  0 |              0 |    2477 | 95.5264   |               2477 |              95.5264 |
|  1 |              1 |     104 |  4.0108   |               2581 |              99.5372 |
|  2 |              2 |      12 |  0.462784 |               2593 |             100      |
+----+----------------+---------+-----------+--------------------+----------------------+ 

+----+----------------+---------+------------+--------------------+----------------------+
|    |   monthly_2g_8 |   count |    percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+------------+--------------------+----------------------|
|  0 |              0 |    2555 | 98.5345    |               2555 |              98.5345 |
|  1 |              1 |      37 |  1.42692   |               2592 |              99.9614 |
|  2 |              2 |       1 |  0.0385654 |               2593 |             100      |
+----+----------------+---------+------------+--------------------+----------------------+ 


Customers who did not churn (Churn : 0)
+----+----------------+---------+------------+--------------------+----------------------+
|    |   monthly_2g_6 |   count |    percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+------------+--------------------+----------------------|
|  0 |              0 |   24228 | 88.3653    |              24228 |              88.3653 |
|  1 |              1 |    2825 | 10.3035    |              27053 |              98.6688 |
|  2 |              2 |     334 |  1.21818   |              27387 |              99.8869 |
|  3 |              3 |      26 |  0.0948282 |              27413 |              99.9818 |
|  4 |              4 |       5 |  0.0182362 |              27418 |             100      |
+----+----------------+---------+------------+--------------------+----------------------+ 

+----+----------------+---------+-------------+--------------------+----------------------+
|    |   monthly_2g_7 |   count |     percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+-------------+--------------------+----------------------|
|  0 |              0 |   24079 | 87.8219     |              24079 |              87.8219 |
|  1 |              1 |    2909 | 10.6098     |              26988 |              98.4317 |
|  2 |              2 |     394 |  1.43701    |              27382 |              99.8687 |
|  3 |              3 |      29 |  0.10577    |              27411 |              99.9745 |
|  4 |              4 |       5 |  0.0182362  |              27416 |              99.9927 |
|  5 |              5 |       2 |  0.00729448 |              27418 |             100      |
+----+----------------+---------+-------------+--------------------+----------------------+ 

+----+----------------+---------+-------------+--------------------+----------------------+
|    |   monthly_2g_8 |   count |     percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+-------------+--------------------+----------------------|
|  0 |              0 |   24383 | 88.9306     |              24383 |              88.9306 |
|  1 |              1 |    2724 |  9.93508    |              27107 |              98.8657 |
|  2 |              2 |     282 |  1.02852    |              27389 |              99.8942 |
|  3 |              3 |      22 |  0.0802393  |              27411 |              99.9745 |
|  4 |              4 |       5 |  0.0182362  |              27416 |              99.9927 |
|  5 |              5 |       2 |  0.00729448 |              27418 |             100      |
+----+----------------+---------+-------------+--------------------+----------------------+

monthly_3g_6, monthly_3g_7, monthly_3g_8¶

columns = ['monthly_3g_6', 'monthly_3g_7', 'monthly_3g_8']
cat_univariate_analysis(columns)

Customers who churned (Churn : 1)
+----+----------------+---------+------------+--------------------+----------------------+
|    |   monthly_3g_6 |   count |    percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+------------+--------------------+----------------------|
|  0 |              0 |    2352 | 90.7057    |               2352 |              90.7057 |
|  1 |              1 |     170 |  6.55611   |               2522 |              97.2619 |
|  2 |              2 |      49 |  1.8897    |               2571 |              99.1516 |
|  3 |              3 |      13 |  0.50135   |               2584 |              99.6529 |
|  4 |              5 |       4 |  0.154261  |               2588 |              99.8072 |
|  5 |              4 |       4 |  0.154261  |               2592 |              99.9614 |
|  6 |              6 |       1 |  0.0385654 |               2593 |             100      |
+----+----------------+---------+------------+--------------------+----------------------+ 

+----+----------------+---------+------------+--------------------+----------------------+
|    |   monthly_3g_7 |   count |    percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+------------+--------------------+----------------------|
|  0 |              0 |    2399 | 92.5183    |               2399 |              92.5183 |
|  1 |              1 |     136 |  5.24489   |               2535 |              97.7632 |
|  2 |              2 |      48 |  1.85114   |               2583 |              99.6143 |
|  3 |              3 |       9 |  0.347088  |               2592 |              99.9614 |
|  4 |              5 |       1 |  0.0385654 |               2593 |             100      |
+----+----------------+---------+------------+--------------------+----------------------+ 

+----+----------------+---------+------------+--------------------+----------------------+
|    |   monthly_3g_8 |   count |    percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+------------+--------------------+----------------------|
|  0 |              0 |    2524 | 97.339     |               2524 |              97.339  |
|  1 |              1 |      56 |  2.15966   |               2580 |              99.4987 |
|  2 |              2 |       8 |  0.308523  |               2588 |              99.8072 |
|  3 |              3 |       4 |  0.154261  |               2592 |              99.9614 |
|  4 |              4 |       1 |  0.0385654 |               2593 |             100      |
+----+----------------+---------+------------+--------------------+----------------------+ 


Customers who did not churn (Churn : 0)
+----+----------------+---------+-------------+--------------------+----------------------+
|    |   monthly_3g_6 |   count |     percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+-------------+--------------------+----------------------|
|  0 |              0 |   24080 | 87.8255     |              24080 |              87.8255 |
|  1 |              1 |    2371 |  8.6476     |              26451 |              96.4731 |
|  2 |              2 |     648 |  2.36341    |              27099 |              98.8365 |
|  3 |              3 |     194 |  0.707564   |              27293 |              99.5441 |
|  4 |              4 |      70 |  0.255307   |              27363 |              99.7994 |
|  5 |              5 |      28 |  0.102123   |              27391 |              99.9015 |
|  6 |              6 |      10 |  0.0364724  |              27401 |              99.938  |
|  7 |              7 |       9 |  0.0328252  |              27410 |              99.9708 |
|  8 |              8 |       3 |  0.0109417  |              27413 |              99.9818 |
|  9 |             11 |       2 |  0.00729448 |              27415 |              99.9891 |
| 10 |              9 |       2 |  0.00729448 |              27417 |              99.9964 |
| 11 |             14 |       1 |  0.00364724 |              27418 |             100      |
+----+----------------+---------+-------------+--------------------+----------------------+ 

+----+----------------+---------+-------------+--------------------+----------------------+
|    |   monthly_3g_7 |   count |     percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+-------------+--------------------+----------------------|
|  0 |              0 |   23962 | 87.3951     |              23962 |              87.3951 |
|  1 |              1 |    2330 |  8.49807    |              26292 |              95.8932 |
|  2 |              2 |     774 |  2.82296    |              27066 |              98.7162 |
|  3 |              3 |     198 |  0.722153   |              27264 |              99.4383 |
|  4 |              4 |      68 |  0.248012   |              27332 |              99.6863 |
|  5 |              5 |      38 |  0.138595   |              27370 |              99.8249 |
|  6 |              6 |      23 |  0.0838865  |              27393 |              99.9088 |
|  7 |              7 |      10 |  0.0364724  |              27403 |              99.9453 |
|  8 |              8 |       5 |  0.0182362  |              27408 |              99.9635 |
|  9 |              9 |       4 |  0.014589   |              27412 |              99.9781 |
| 10 |             11 |       2 |  0.00729448 |              27414 |              99.9854 |
| 11 |             16 |       1 |  0.00364724 |              27415 |              99.9891 |
| 12 |             14 |       1 |  0.00364724 |              27416 |              99.9927 |
| 13 |             12 |       1 |  0.00364724 |              27417 |              99.9964 |
| 14 |             10 |       1 |  0.00364724 |              27418 |             100      |
+----+----------------+---------+-------------+--------------------+----------------------+ 

+----+----------------+---------+-------------+--------------------+----------------------+
|    |   monthly_3g_8 |   count |     percent |   cumulative_count |   cumulative_percent |
|----+----------------+---------+-------------+--------------------+----------------------|
|  0 |              0 |   24002 | 87.541      |              24002 |              87.541  |
|  1 |              1 |    2347 |  8.56007    |              26349 |              96.1011 |
|  2 |              2 |     728 |  2.65519    |              27077 |              98.7563 |
|  3 |              3 |     193 |  0.703917   |              27270 |              99.4602 |
|  4 |              4 |      86 |  0.313663   |              27356 |              99.7739 |
|  5 |              5 |      30 |  0.109417   |              27386 |              99.8833 |
|  6 |              6 |      14 |  0.0510613  |              27400 |              99.9343 |
|  7 |              7 |       9 |  0.0328252  |              27409 |              99.9672 |
|  8 |              9 |       3 |  0.0109417  |              27412 |              99.9781 |
|  9 |              8 |       3 |  0.0109417  |              27415 |              99.9891 |
| 10 |             10 |       2 |  0.00729448 |              27417 |              99.9964 |
| 11 |             16 |       1 |  0.00364724 |              27418 |             100      |
+----+----------------+---------+-------------+--------------------+----------------------+

sachet_3g_6, sachet_3g_7, sachet_3g_8¶

columns = ['sachet_3g_6', 'sachet_3g_7','sachet_3g_8']
print(data[columns].dtypes)
cat_univariate_analysis(columns)

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-1-43696c0d750b> in <module>
      1 columns = ['sachet_3g_6', 'sachet_3g_7','sachet_3g_8']
----> 2 print(data[columns].dtypes)
      3 cat_univariate_analysis(columns)

NameError: name 'data' is not defined

aug_vbc_3g, jul_vbc_3g, jun_vbc_3g¶

columns = [ 'vbc_3g_6', 'vbc_3g_7','vbc_3g_8']
num_univariate_analysis(columns, 'log')

Customers who churned (Churn : 1)
          vbc_3g_6     vbc_3g_7     vbc_3g_8
count  2593.000000  2593.000000  2593.000000
mean     81.564601    71.143880    32.610659
std     320.898511   284.882601   197.998246
min       0.000000     0.000000     0.000000
25%       0.000000     0.000000     0.000000
50%       0.000000     0.000000     0.000000
75%       0.000000     0.000000     0.000000
max    6931.810000  4908.270000  5738.740000

Customers who did not churn (Churn : 0)
           vbc_3g_6      vbc_3g_7      vbc_3g_8
count  27418.000000  27418.000000  27418.000000
mean     125.124167    141.178182    138.597023
std      395.413666    417.292310    402.761779
min        0.000000      0.000000      0.000000
25%        0.000000      0.000000      0.000000
50%        0.000000      0.000000      0.000000
75%        0.000000      9.940000     17.675000
max    11166.210000   9165.600000  12916.220000

Bivariate Analysis¶

data.head()

'total_og_mou_6' vs 'total_og_mou_8' with respect to Churn.¶

sns.scatterplot(x=data['total_og_mou_6'],y=data['total_og_mou_8'],hue=data['Churn'])

<matplotlib.axes._subplots.AxesSubplot at 0x7ffc15cc7190>

'total_og_mou_7' vs 'total_og_mou_8' with respect to Churn.¶

sns.scatterplot(x=data['total_og_mou_6'],y=data['total_og_mou_8'],hue=data['Churn'])

<matplotlib.axes._subplots.AxesSubplot at 0x7ffc1a6f59a0>

The customers with lower total_og_mou in 6th and 8th months are more likely to Churn compared to the ones with higher total_og_mou.

'aon' vs 'total_og_mou_8' with respect to Churn.¶

sns.scatterplot(x=data['aon'],y=data['total_og_mou_8'],hue=data['Churn'])

<matplotlib.axes._subplots.AxesSubplot at 0x7ffc128bd790>

The customers with lesser total_og_mou_8 and aon are more likely to churn compared to the one with higher total_og_mou_8 and aon.

sns.scatterplot(x=data['aon'],y=data['total_ic_mou_8'],hue=data['Churn'])

<matplotlib.axes._subplots.AxesSubplot at 0x7ffc197fbdc0>

The customers with less total_ic_mou_8 are more likely to churn irrespective of aon.
The customers with total_ic_mou_8 > 2000 are very less likely to churn.

'max_rech_amt_6' vs 'max_rech_amt_8' with respect to 'Churn'.¶

sns.scatterplot(x=data['max_rech_amt_6'],y=data['max_rech_amt_8'],hue=data['Churn'])

<matplotlib.axes._subplots.AxesSubplot at 0x7ffc1b2ad970>

Correlation Analysis¶

# function to correlate variables
def correlation(dataframe) : 
    
    columnsForAnalysis = set(dataframe.columns.values) - {'Churn'}
    cor0=dataframe[columnsForAnalysis].corr()
    type(cor0)
    cor0.where(np.triu(np.ones(cor0.shape),k=1).astype(np.bool))
    cor0=cor0.unstack().reset_index()
    cor0.columns=['VAR1','VAR2','CORR']
    cor0.dropna(subset=['CORR'], inplace=True)
    cor0.CORR=round(cor0['CORR'],2)
    cor0.CORR=cor0.CORR.abs()
    cor0.sort_values(by=['CORR'],ascending=False)
    cor0=cor0[~(cor0['VAR1']==cor0['VAR2'])]

    # removing duplicate correlations
    cor0['pair'] = cor0[['VAR1', 'VAR2']].apply(lambda x: '{}-{}'.format(*sorted((x[0], x[1]))), axis=1)
    
    cor0 = cor0.drop_duplicates(subset=['pair'], keep='first')
    cor0 = cor0[['VAR1', 'VAR2','CORR']]
    return pd.DataFrame(cor0.sort_values(by=['CORR'],ascending=False))

# Correlations for Churn : 0  - non churn customers
# Absolute values are reported 
pd.set_option('precision', 2)
cor_0 = correlation(non_churned_customers)

# filtering for correlations >= 40%
condition = cor_0['CORR'] > 0.4
cor_0 = cor_0[condition]
cor_0.style.background_gradient(cmap='GnBu').hide_index()

# Correlations for Churn : 1  - churned customers
# Absolute values are reported 
pd.set_option('precision', 2)
cor_1 = correlation(churned_customers)

# filtering for correlations >= 40%
condition = cor_1['CORR'] > 0.4
cor_1 = cor_1[condition]
cor_1.style.background_gradient(cmap='GnBu').hide_index()

Data Preparation¶

Derived Variables¶

# Derived variables to measure change in usage 

# Usage 
data['delta_vol_2g'] = data['vol_2g_mb_8'] - data['vol_2g_mb_6'].add(data['vol_2g_mb_7']).div(2)
data['delta_vol_3g'] = data['vol_3g_mb_8'] - data['vol_3g_mb_6'].add(data['vol_3g_mb_7']).div(2)
data['delta_total_og_mou'] = data['total_og_mou_8'] - data['total_og_mou_6'].add(data['total_og_mou_7']).div(2)
data['delta_total_ic_mou'] = data['total_ic_mou_8'] - data['total_ic_mou_6'].add(data['total_ic_mou_7']).div(2)
data['delta_vbc_3g'] = data['vbc_3g_8'] - data['vbc_3g_6'].add(data['vbc_3g_7']).div(2)

# Revenue 
data['delta_arpu'] = data['arpu_8'] - data['arpu_6'].add(data['arpu_7']).div(2)
data['delta_total_rech_amt'] = data['total_rech_amt_8'] - data['total_rech_amt_6'].add(data['total_rech_amt_7']).div(2)

# Removing variables used for derivation : 
data.drop(columns=[
 'vol_2g_mb_8', 'vol_2g_mb_6', 'vol_2g_mb_7',
  'vol_3g_mb_8'  , 'vol_3g_mb_6', 'vol_3g_mb_7' ,
    'total_og_mou_8','total_og_mou_6', 'total_og_mou_7', 
    'total_ic_mou_8','total_ic_mou_6', 'total_ic_mou_7',
    'vbc_3g_8','vbc_3g_6','vbc_3g_7',
    'arpu_8','arpu_6','arpu_7',
    'total_rech_amt_8', 'total_rech_amt_6', 'total_rech_amt_7'
    
], inplace=True)

Outlier Treatment¶

# Looking at quantiles from 0.90 to 1. 
data.quantile(np.arange(0.9,1.01,0.01)).style.bar()

# Looking at percentage change in quantiles from 0.90 to 1. 
data.quantile(np.arange(0.9,1.01,0.01)).pct_change().mul(100).style.bar()

# Columns with outliers 
pct_change_99_1 = data.quantile(np.arange(0.9,1.01,0.01)).pct_change().mul(100).iloc[-1]
outlier_condition = pct_change_99_1 > 100
columns_with_outliers = pct_change_99_1[outlier_condition].index.values
print('Columns with outliers :\n', columns_with_outliers)

Columns with outliers :
 ['onnet_mou_6' 'onnet_mou_7' 'onnet_mou_8' 'offnet_mou_6' 'offnet_mou_7'
 'offnet_mou_8' 'roam_ic_mou_6' 'roam_ic_mou_7' 'roam_ic_mou_8'
 'roam_og_mou_6' 'roam_og_mou_7' 'roam_og_mou_8' 'loc_og_t2t_mou_6'
 'loc_og_t2t_mou_7' 'loc_og_t2t_mou_8' 'loc_og_t2m_mou_6'
 'loc_og_t2m_mou_7' 'loc_og_t2m_mou_8' 'loc_og_t2f_mou_6'
 'loc_og_t2f_mou_7' 'loc_og_t2f_mou_8' 'loc_og_t2c_mou_6'
 'loc_og_t2c_mou_7' 'loc_og_t2c_mou_8' 'loc_og_mou_6' 'loc_og_mou_7'
 'loc_og_mou_8' 'std_og_t2t_mou_6' 'std_og_t2t_mou_7' 'std_og_t2t_mou_8'
 'std_og_t2m_mou_6' 'std_og_t2m_mou_7' 'std_og_t2m_mou_8'
 'std_og_t2f_mou_6' 'std_og_t2f_mou_7' 'std_og_t2f_mou_8' 'std_og_mou_6'
 'std_og_mou_7' 'std_og_mou_8' 'isd_og_mou_6' 'isd_og_mou_7'
 'isd_og_mou_8' 'spl_og_mou_6' 'spl_og_mou_7' 'spl_og_mou_8' 'og_others_6'
 'og_others_7' 'og_others_8' 'loc_ic_t2t_mou_6' 'loc_ic_t2t_mou_7'
 'loc_ic_t2t_mou_8' 'loc_ic_t2m_mou_6' 'loc_ic_t2m_mou_7'
 'loc_ic_t2m_mou_8' 'loc_ic_t2f_mou_6' 'loc_ic_t2f_mou_7'
 'loc_ic_t2f_mou_8' 'loc_ic_mou_6' 'loc_ic_mou_7' 'loc_ic_mou_8'
 'std_ic_t2t_mou_6' 'std_ic_t2t_mou_7' 'std_ic_t2t_mou_8'
 'std_ic_t2m_mou_6' 'std_ic_t2m_mou_7' 'std_ic_t2m_mou_8'
 'std_ic_t2f_mou_6' 'std_ic_t2f_mou_7' 'std_ic_t2f_mou_8' 'std_ic_mou_6'
 'std_ic_mou_7' 'std_ic_mou_8' 'spl_ic_mou_6' 'spl_ic_mou_7'
 'spl_ic_mou_8' 'isd_ic_mou_6' 'isd_ic_mou_7' 'isd_ic_mou_8' 'ic_others_6'
 'ic_others_7' 'ic_others_8' 'total_rech_num_6' 'total_rech_num_7'
 'total_rech_num_8' 'max_rech_amt_6' 'max_rech_amt_7' 'max_rech_amt_8'
 'last_day_rch_amt_6' 'last_day_rch_amt_7' 'last_day_rch_amt_8'
 'Average_rech_amt_6n7' 'delta_vol_2g' 'delta_vol_3g' 'delta_total_og_mou'
 'delta_total_ic_mou' 'delta_vbc_3g' 'delta_arpu' 'delta_total_rech_amt']

# capping outliers to 99th percentile values
outlier_treatment = pd.DataFrame(columns=['Column', 'Outlier Threshold', 'Outliers replaced'])
for col in columns_with_outliers : 
    outlier_threshold = data[col].quantile(0.99)
    condition = data[col] > outlier_threshold
    outlier_treatment = outlier_treatment.append({'Column' : col , 'Outlier Threshold' : outlier_threshold, 'Outliers replaced' : data.loc[condition,col].shape[0] }, ignore_index=True)
    data.loc[condition, col] = outlier_threshold
outlier_treatment

categorical = data.dtypes == 'category'
categorical_vars = data.columns[categorical].to_list()
ind_categorical_vars = set(categorical_vars) - {'Churn'} #independent categorical variables
ind_categorical_vars

{'monthly_2g_6',
 'monthly_2g_7',
 'monthly_2g_8',
 'monthly_3g_6',
 'monthly_3g_7',
 'monthly_3g_8',
 'sachet_2g_6',
 'sachet_2g_7',
 'sachet_2g_8',
 'sachet_3g_6',
 'sachet_3g_7',
 'sachet_3g_8'}

Grouping Categories with less Contribution¶

# Finding & Grouping categories with less than 1% contribution in each column into "Others"
for col in ind_categorical_vars : 
    category_counts = 100*data[col].value_counts(normalize=True)
    print('\n',tabulate(pd.DataFrame(category_counts), headers='keys', tablefmt='psql'),'\n')
    low_count_categories = category_counts[category_counts <= 1].index.to_list()
    print(f"Replaced {low_count_categories} in {col} with category : Others")
    data[col].replace(low_count_categories,'Others',inplace=True)

 +----+---------------+
|    |   sachet_3g_6 |
|----+---------------|
|  0 |   93.4091     |
|  1 |    4.35507    |
|  2 |    1.04295    |
|  3 |    0.396521   |
|  4 |    0.219919   |
|  5 |    0.123288   |
|  6 |    0.089967   |
|  7 |    0.0866349  |
|  8 |    0.0499817  |
|  9 |    0.0499817  |
| 10 |    0.0366532  |
| 11 |    0.0266569  |
| 15 |    0.0166606  |
| 12 |    0.0133284  |
| 19 |    0.0133284  |
| 13 |    0.00999633 |
| 14 |    0.00999633 |
| 18 |    0.00999633 |
| 23 |    0.00999633 |
| 16 |    0.00666422 |
| 22 |    0.00666422 |
| 29 |    0.00666422 |
| 28 |    0.00333211 |
| 17 |    0.00333211 |
| 21 |    0.00333211 |
+----+---------------+ 

Replaced [3, 4, 5, 6, 7, 8, 9, 10, 11, 15, 12, 19, 13, 14, 18, 23, 16, 22, 29, 28, 17, 21] in sachet_3g_6 with category : Others

 +----+----------------+
|    |   monthly_2g_7 |
|----+----------------|
|  0 |    88.4876     |
|  1 |    10.0397     |
|  2 |     1.35284    |
|  3 |     0.0966312  |
|  4 |     0.0166606  |
|  5 |     0.00666422 |
+----+----------------+ 

Replaced [3, 4, 5] in monthly_2g_7 with category : Others

 +----+----------------+
|    |   monthly_2g_8 |
|----+----------------|
|  0 |    89.7604     |
|  1 |     9.19996    |
|  2 |     0.942988   |
|  3 |     0.0733065  |
|  4 |     0.0166606  |
|  5 |     0.00666422 |
+----+----------------+ 

Replaced [2, 3, 4, 5] in monthly_2g_8 with category : Others

 +----+---------------+
|    |   sachet_3g_8 |
|----+---------------|
|  0 |   94.2388     |
|  1 |    3.52537    |
|  2 |    0.839692   |
|  3 |    0.429842   |
|  4 |    0.243244   |
|  5 |    0.219919   |
|  6 |    0.0866349  |
|  7 |    0.0766386  |
|  8 |    0.0733065  |
|  9 |    0.0399853  |
| 12 |    0.0366532  |
| 13 |    0.0333211  |
| 10 |    0.0333211  |
| 11 |    0.0199927  |
| 14 |    0.0199927  |
| 15 |    0.0166606  |
| 16 |    0.00999633 |
| 17 |    0.00666422 |
| 18 |    0.00666422 |
| 20 |    0.00666422 |
| 21 |    0.00666422 |
| 23 |    0.00666422 |
| 38 |    0.00333211 |
| 19 |    0.00333211 |
| 25 |    0.00333211 |
| 27 |    0.00333211 |
| 29 |    0.00333211 |
| 30 |    0.00333211 |
| 41 |    0.00333211 |
+----+---------------+ 

Replaced [2, 3, 4, 5, 6, 7, 8, 9, 12, 13, 10, 11, 14, 15, 16, 17, 18, 20, 21, 23, 38, 19, 25, 27, 29, 30, 41] in sachet_3g_8 with category : Others

 +----+----------------+
|    |   monthly_3g_7 |
|----+----------------|
|  0 |    87.8378     |
|  1 |     8.21699    |
|  2 |     2.739      |
|  3 |     0.689747   |
|  4 |     0.226584   |
|  5 |     0.129952   |
|  6 |     0.0766386  |
|  7 |     0.0333211  |
|  8 |     0.0166606  |
|  9 |     0.0133284  |
| 11 |     0.00666422 |
| 16 |     0.00333211 |
| 14 |     0.00333211 |
| 12 |     0.00333211 |
| 10 |     0.00333211 |
+----+----------------+ 

Replaced [3, 4, 5, 6, 7, 8, 9, 11, 16, 14, 12, 10] in monthly_3g_7 with category : Others

 +----+---------------+
|    |   sachet_2g_6 |
|----+---------------|
|  0 |   82.5631     |
|  1 |    7.87378    |
|  2 |    3.3621     |
|  3 |    2.0126     |
|  4 |    1.32951    |
|  5 |    0.703076   |
|  6 |    0.509813   |
|  7 |    0.356536   |
|  8 |    0.286562   |
|  9 |    0.239912   |
| 10 |    0.17327    |
| 12 |    0.146613   |
| 11 |    0.0999633  |
| 13 |    0.0566459  |
| 14 |    0.0533138  |
| 15 |    0.0433175  |
| 17 |    0.0366532  |
| 18 |    0.029989   |
| 19 |    0.029989   |
| 16 |    0.0233248  |
| 22 |    0.0133284  |
| 20 |    0.00999633 |
| 21 |    0.00999633 |
| 24 |    0.00999633 |
| 25 |    0.00999633 |
| 39 |    0.00333211 |
| 27 |    0.00333211 |
| 30 |    0.00333211 |
| 32 |    0.00333211 |
| 34 |    0.00333211 |
| 28 |    0          |
| 42 |    0          |
+----+---------------+ 

Replaced [5, 6, 7, 8, 9, 10, 12, 11, 13, 14, 15, 17, 18, 19, 16, 22, 20, 21, 24, 25, 39, 27, 30, 32, 34, 28, 42] in sachet_2g_6 with category : Others

 +----+----------------+
|    |   monthly_2g_6 |
|----+----------------|
|  0 |     88.9074    |
|  1 |      9.83306   |
|  2 |      1.14958   |
|  3 |      0.0866349 |
|  4 |      0.0233248 |
+----+----------------+ 

Replaced [3, 4] in monthly_2g_6 with category : Others

 +----+---------------+
|    |   sachet_2g_7 |
|----+---------------|
|  0 |   81.8033     |
|  1 |    7.24068    |
|  2 |    3.34877    |
|  3 |    1.96595    |
|  4 |    1.50945    |
|  5 |    1.20622    |
|  6 |    0.843024   |
|  7 |    0.543134   |
|  8 |    0.403185   |
| 10 |    0.239912   |
|  9 |    0.219919   |
| 11 |    0.159941   |
| 12 |    0.0966312  |
| 14 |    0.0799707  |
| 13 |    0.0666422  |
| 15 |    0.0499817  |
| 16 |    0.0366532  |
| 18 |    0.0333211  |
| 17 |    0.029989   |
| 20 |    0.0266569  |
| 19 |    0.0233248  |
| 21 |    0.00999633 |
| 26 |    0.00999633 |
| 27 |    0.00999633 |
| 22 |    0.00666422 |
| 23 |    0.00666422 |
| 30 |    0.00666422 |
| 42 |    0.00333211 |
| 24 |    0.00333211 |
| 25 |    0.00333211 |
| 29 |    0.00333211 |
| 32 |    0.00333211 |
| 35 |    0.00333211 |
| 48 |    0.00333211 |
| 28 |    0          |
+----+---------------+ 

Replaced [6, 7, 8, 10, 9, 11, 12, 14, 13, 15, 16, 18, 17, 20, 19, 21, 26, 27, 22, 23, 30, 42, 24, 25, 29, 32, 35, 48, 28] in sachet_2g_7 with category : Others

 +----+---------------+
|    |   sachet_3g_7 |
|----+---------------|
|  0 |   93.4757     |
|  1 |    4.10849    |
|  2 |    1.03962    |
|  3 |    0.383193   |
|  4 |    0.239912   |
|  5 |    0.219919   |
|  6 |    0.139949   |
|  7 |    0.059978   |
|  9 |    0.0533138  |
|  8 |    0.0466496  |
| 11 |    0.0433175  |
| 10 |    0.0333211  |
| 12 |    0.0333211  |
| 15 |    0.0166606  |
| 14 |    0.0166606  |
| 13 |    0.0133284  |
| 18 |    0.0133284  |
| 19 |    0.00999633 |
| 20 |    0.00999633 |
| 22 |    0.00999633 |
| 17 |    0.00666422 |
| 21 |    0.00666422 |
| 24 |    0.00666422 |
| 33 |    0.00333211 |
| 16 |    0.00333211 |
| 31 |    0.00333211 |
| 35 |    0.00333211 |
+----+---------------+ 

Replaced [3, 4, 5, 6, 7, 9, 8, 11, 10, 12, 15, 14, 13, 18, 19, 20, 22, 17, 21, 24, 33, 16, 31, 35] in sachet_3g_7 with category : Others

 +----+----------------+
|    |   monthly_3g_8 |
|----+----------------|
|  0 |    88.3876     |
|  1 |     8.00706    |
|  2 |     2.45243    |
|  3 |     0.656426   |
|  4 |     0.289894   |
|  5 |     0.0999633  |
|  6 |     0.0466496  |
|  7 |     0.029989   |
|  9 |     0.00999633 |
|  8 |     0.00999633 |
| 10 |     0.00666422 |
| 16 |     0.00333211 |
+----+----------------+ 

Replaced [3, 4, 5, 6, 7, 9, 8, 10, 16] in monthly_3g_8 with category : Others

 +----+----------------+
|    |   monthly_3g_6 |
|----+----------------|
|  0 |    88.0744     |
|  1 |     8.4669     |
|  2 |     2.32248    |
|  3 |     0.689747   |
|  4 |     0.246576   |
|  5 |     0.106628   |
|  6 |     0.0366532  |
|  7 |     0.029989   |
|  8 |     0.00999633 |
| 11 |     0.00666422 |
|  9 |     0.00666422 |
| 14 |     0.00333211 |
+----+----------------+ 

Replaced [3, 4, 5, 6, 7, 8, 11, 9, 14] in monthly_3g_6 with category : Others

 +----+---------------+
|    |   sachet_2g_8 |
|----+---------------|
|  0 |   79.7274     |
|  1 |    8.87008    |
|  2 |    3.25881    |
|  3 |    2.19253    |
|  4 |    1.81267    |
|  5 |    1.44947    |
|  6 |    0.88301    |
|  7 |    0.459831   |
|  8 |    0.313218   |
|  9 |    0.249908   |
| 10 |    0.169938   |
| 11 |    0.123288   |
| 12 |    0.113292   |
| 14 |    0.0766386  |
| 15 |    0.0566459  |
| 13 |    0.0499817  |
| 16 |    0.0433175  |
| 18 |    0.0266569  |
| 17 |    0.0233248  |
| 19 |    0.0233248  |
| 20 |    0.0133284  |
| 34 |    0.00666422 |
| 29 |    0.00666422 |
| 27 |    0.00666422 |
| 24 |    0.00666422 |
| 22 |    0.00666422 |
| 21 |    0.00666422 |
| 23 |    0.00333211 |
| 25 |    0.00333211 |
| 26 |    0.00333211 |
| 31 |    0.00333211 |
| 32 |    0.00333211 |
| 33 |    0.00333211 |
| 44 |    0.00333211 |
+----+---------------+ 

Replaced [6, 7, 8, 9, 10, 11, 12, 14, 15, 13, 16, 18, 17, 19, 20, 34, 29, 27, 24, 22, 21, 23, 25, 26, 31, 32, 33, 44] in sachet_2g_8 with category : Others

Creating Dummy Variables¶

dummy_vars = pd.get_dummies(data[ind_categorical_vars], drop_first=False, prefix=ind_categorical_vars, prefix_sep='_')
dummy_vars.head()

reference_cols = dummy_vars.filter(regex='.*Others$').columns.to_list() # Using category 'Others' in each column as reference. 
dummy_vars.drop(columns=reference_cols, inplace=True)
reference_cols

['sachet_3g_6_Others',
 'monthly_2g_7_Others',
 'monthly_2g_8_Others',
 'sachet_3g_8_Others',
 'monthly_3g_7_Others',
 'sachet_2g_6_Others',
 'monthly_2g_6_Others',
 'sachet_2g_7_Others',
 'sachet_3g_7_Others',
 'monthly_3g_8_Others',
 'monthly_3g_6_Others',
 'sachet_2g_8_Others']

# concatenating dummy variables with original 'data'
data.drop(columns=ind_categorical_vars, inplace=True) # dropping original categorical columns
data = pd.concat([data, dummy_vars], axis=1)
data.head()

dummy_cols = dummy_vars.columns.to_list()
data[dummy_cols] = data[dummy_cols].astype('category')

data.shape

(30011, 142)

----joint

This following section contains

Test Train Split
Class Imbalance
Standardization
Modelling
- Model 1 : Logistic Regression with RFE & Manual Elimination ( Interpretable Model )
- Model 2 : PCA + Logistic Regression
- Model 3 : PCA + Random Forest Classifier
- Model 4 : PCA + XGBoost

Train-Test Split¶

y = data.pop('Churn') # Predicted / Target Variable
X = data # Predictor variables

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X,y, train_size=0.7, random_state=42)

Class Imbalance¶

y.value_counts(normalize=True).to_frame()

# Ratio of classes 
class_0 = y[y == 0].count()
class_1 = y[y == 1].count()

print(f'Class Imbalance Ratio : {round(class_1/class_0,3)}')

Class Imbalance Ratio : 0.095

To account for class imbalance, Synthetic Minority Class Oversampling Technique (SMOTE) could be used.

Using SMOTE¶

#!pip install imblearn
from imblearn.over_sampling import SMOTE
smt = SMOTE(random_state=42, k_neighbors=5)

# Resampling Train set to account for class imbalance

X_train_resampled, y_train_resampled= smt.fit_resample(X_train, y_train)
X_train_resampled.head()

Standardizing Columns¶

# columns with numerical data
condition1 = data.dtypes == 'int'
condition2 = data.dtypes == 'float'
numerical_vars = data.columns[condition1 | condition2].to_list()

# Standard scaling
from sklearn.preprocessing import StandardScaler
scaler = StandardScaler() 

# Fit and transform train set 
X_train_resampled[numerical_vars] = scaler.fit_transform(X_train_resampled[numerical_vars])

# Transform test set
X_test[numerical_vars] = scaler.transform(X_test[numerical_vars])

# summary statistics of standardized variables
round(X_train_resampled.describe(),2)

Modelling¶

Model 1 : Interpretable Model : Logistic Regression¶

Baseline Logistic Regression Model¶

from sklearn.linear_model import LogisticRegression


baseline_model = LogisticRegression(random_state=100, class_weight='balanced') # `weight of class` balancing technique used
baseline_model = baseline_model.fit(X_train, y_train)

y_train_pred = baseline_model.predict_proba(X_train)[:,1]
y_test_pred  = baseline_model.predict_proba(X_test)[:,1]

y_train_pred = pd.Series(y_train_pred,index = X_train.index, ) # converting test and train to a series to preserve index
y_test_pred = pd.Series(y_test_pred,index = X_test.index)

Baseline Performance

# Function for Baseline Performance Metrics
import math
def model_metrics(matrix) :
    TN = matrix[0][0]
    TP = matrix[1][1]
    FP = matrix[0][1]
    FN = matrix[1][0]
    accuracy = round((TP + TN)/float(TP+TN+FP+FN),3)
    print('Accuracy :' ,accuracy )
    sensitivity = round(TP/float(FN + TP),3)
    print('Sensitivity / True Positive Rate / Recall :', sensitivity)
    specificity = round(TN/float(TN + FP),3)
    print('Specificity / True Negative Rate : ', specificity)
    precision = round(TP/float(TP + FP),3)
    print('Precision / Positive Predictive Value :', precision)
    print('F1-score :', round(2*precision*sensitivity/(precision + sensitivity),3))

# Prediction at threshold of 0.5 
classification_threshold = 0.5 
    
y_train_pred_classified = y_train_pred.map(lambda x : 1 if x > classification_threshold else 0)
y_test_pred_classified = y_test_pred.map(lambda x : 1 if x > classification_threshold else 0)

from sklearn.metrics import confusion_matrix
train_matrix = confusion_matrix(y_train, y_train_pred_classified)
print('Confusion Matrix for train:\n', train_matrix)
test_matrix = confusion_matrix(y_test, y_test_pred_classified)
print('\nConfusion Matrix for test: \n', test_matrix)

Confusion Matrix for train:
 [[16001  3186]
 [  326  1494]]

Confusion Matrix for test: 
 [[6090 2141]
 [ 149  624]]

# Baseline Model Performance : 

print('Train Performance : \n')
model_metrics(train_matrix)

print('\n\nTest Performance : \n')
model_metrics(test_matrix)

Train Performance : 

Accuracy : 0.833
Sensitivity / True Positive Rate / Recall : 0.821
Specificity / True Negative Rate :  0.834
Precision / Positive Predictive Value : 0.319
F1-score : 0.459


Test Performance : 

Accuracy : 0.746
Sensitivity / True Positive Rate / Recall : 0.807
Specificity / True Negative Rate :  0.74
Precision / Positive Predictive Value : 0.226
F1-score : 0.353

Baseline Performance - Finding Optimum Probability Cutoff

# Specificity / Sensitivity Tradeoff 

# Classification at probability thresholds between 0 and 1 
y_train_pred_thres = pd.DataFrame(index=X_train.index)
thresholds = [float(x)/10 for x in range(10)]

def thresholder(x, thresh) :
    if x > thresh : 
        return 1 
    else : 
        return 0

    
for i in thresholds:
    y_train_pred_thres[i]= y_train_pred.map(lambda x : thresholder(x,i))
y_train_pred_thres.head()

# # sensitivity, specificity, accuracy for each threshold
metrics_df = pd.DataFrame(columns=['sensitivity', 'specificity', 'accuracy'])

# Function for calculation of metrics for each threshold
def model_metrics_thres(matrix) :
    TN = matrix[0][0]
    TP = matrix[1][1]
    FP = matrix[0][1]
    FN = matrix[1][0]
    accuracy = round((TP + TN)/float(TP+TN+FP+FN),3)
    sensitivity = round(TP/float(FN + TP),3)
    specificity = round(TN/float(TN + FP),3)
    return sensitivity,specificity,accuracy

# generating a data frame for metrics for each threshold
for thres,column in zip(thresholds,y_train_pred_thres.columns.to_list()) : 
    confusion = confusion_matrix(y_train, y_train_pred_thres.loc[:,column])
    sensitivity,specificity,accuracy = model_metrics_thres(confusion)
    
    metrics_df =  metrics_df.append({ 
        'sensitivity' :sensitivity,
        'specificity' : specificity,
        'accuracy' : accuracy
    }, ignore_index = True)
    
metrics_df.index = thresholds
metrics_df

metrics_df.plot(kind='line', figsize=(24,8), grid=True, xticks=np.arange(0,1,0.02),
                title='Specificity-Sensitivity TradeOff');

Baseline Performance at Optimum Cutoff¶

optimum_cutoff = 0.49
y_train_pred_final = y_train_pred.map(lambda x : 1 if x > optimum_cutoff else 0)
y_test_pred_final = y_test_pred.map(lambda x : 1 if x > optimum_cutoff else 0)

train_matrix = confusion_matrix(y_train, y_train_pred_final)
print('Confusion Matrix for train:\n', train_matrix)
test_matrix = confusion_matrix(y_test, y_test_pred_final)
print('\nConfusion Matrix for test: \n', test_matrix)

Confusion Matrix for train:
 [[15888  3299]
 [  318  1502]]

Confusion Matrix for test: 
 [[1329 6902]
 [  16  757]]

print('Train Performance: \n')
model_metrics(train_matrix)

print('\n\nTest Performance : \n')
model_metrics(test_matrix)

Train Performance: 

Accuracy : 0.828
Sensitivity / True Positive Rate / Recall : 0.825
Specificity / True Negative Rate :  0.828
Precision / Positive Predictive Value : 0.313
F1-score : 0.454


Test Performance : 

Accuracy : 0.232
Sensitivity / True Positive Rate / Recall : 0.979
Specificity / True Negative Rate :  0.161
Precision / Positive Predictive Value : 0.099
F1-score : 0.18

# ROC_AUC score 
from sklearn.metrics import roc_auc_score
print('ROC AUC score for Train : ',round(roc_auc_score(y_train, y_train_pred),3), '\n' )
print('ROC AUC score for Test : ',round(roc_auc_score(y_test, y_test_pred),3) )

ROC AUC score for Train :  0.891 

ROC AUC score for Test :  0.838

Feature Selection using RFE¶

from sklearn.feature_selection import RFE
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression(random_state=100 , class_weight='balanced')
rfe = RFE(lr, 15)
results = rfe.fit(X_train,y_train)
results.support_

array([False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False,  True,
       False, False, False, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
       False, False,  True, False, False, False, False, False, False,
       False, False, False, False,  True,  True, False, False, False,
       False, False, False, False, False, False, False, False, False,
        True, False,  True, False, False, False, False, False, False,
       False, False, False, False, False, False, False, False, False,
        True, False, False,  True, False, False,  True, False, False,
       False,  True, False, False,  True, False, False, False, False,
        True, False, False,  True, False, False, False, False, False,
       False, False, False,  True, False, False, False, False, False,
        True, False, False, False, False, False])

# DataFrame with features supported by RFE
rfe_support = pd.DataFrame({'Column' : X.columns.to_list(), 'Rank' : rfe.ranking_, 
                                      'Support' :  rfe.support_}).sort_values(by=
                                       'Rank', ascending=True)
rfe_support

# RFE Selected columns
rfe_selected_columns = rfe_support.loc[rfe_support['Rank'] == 1,'Column'].to_list()
rfe_selected_columns

['sachet_3g_6_0',
 'sachet_2g_7_0',
 'monthly_2g_7_0',
 'sachet_2g_8_0',
 'total_rech_num_6',
 'monthly_3g_8_0',
 'monthly_2g_8_0',
 'total_rech_num_8',
 'monthly_2g_6_0',
 'std_ic_t2f_mou_8',
 'std_ic_t2f_mou_7',
 'sachet_2g_6_0',
 'monthly_3g_7_0',
 'loc_ic_t2f_mou_8',
 'std_og_t2f_mou_8']

Logistic Regression with RFE Selected Columns¶

Model I¶

# Logistic Regression Model with RFE columns
import statsmodels.api as sm 

# Note that the SMOTE resampled Train set is used with statsmodels.api.GLM since it doesnot support class_weight
logr = sm.GLM(y_train_resampled,(sm.add_constant(X_train_resampled[rfe_selected_columns])), family = sm.families.Binomial())
logr_fit = logr.fit()
logr_fit.summary()

Logistic Regression with Manual Feature Elimination¶

# Using P-value and vif for manual feature elimination

from statsmodels.stats.outliers_influence import variance_inflation_factor
def vif(X_train_resampled, logr_fit, selected_columns) : 
    vif = pd.DataFrame()
    vif['Features'] = rfe_selected_columns
    vif['VIF'] = [variance_inflation_factor(X_train_resampled[selected_columns].values, i) for i in range(X_train_resampled[selected_columns].shape[1])]
    vif['VIF'] = round(vif['VIF'], 2)
    vif = vif.set_index('Features')
    vif['P-value'] = round(logr_fit.pvalues,4)
    vif = vif.sort_values(by = ["VIF",'P-value'], ascending = [False,False])
    return vif

vif(X_train_resampled, logr_fit, rfe_selected_columns)

'monthly_2g_7_0' has the very p-value. Hence, this feature could be eliminated

selected_columns = rfe_selected_columns
selected_columns.remove('monthly_2g_7_0')
selected_columns

['sachet_3g_6_0',
 'sachet_2g_7_0',
 'sachet_2g_8_0',
 'total_rech_num_6',
 'monthly_3g_8_0',
 'monthly_2g_8_0',
 'total_rech_num_8',
 'monthly_2g_6_0',
 'std_ic_t2f_mou_8',
 'std_ic_t2f_mou_7',
 'sachet_2g_6_0',
 'monthly_3g_7_0',
 'loc_ic_t2f_mou_8',
 'std_og_t2f_mou_8']

Model II¶

logr2 = sm.GLM(y_train_resampled,(sm.add_constant(X_train_resampled[selected_columns])), family = sm.families.Binomial())
logr2_fit = logr2.fit()
logr2_fit.summary()

# vif and p-values
vif(X_train_resampled, logr2_fit, selected_columns)

'monthly_2g_6_0' has very high p-value. Hence, this feature could be eliminated

selected_columns.remove('monthly_2g_6_0')
selected_columns

['sachet_3g_6_0',
 'sachet_2g_7_0',
 'sachet_2g_8_0',
 'total_rech_num_6',
 'monthly_3g_8_0',
 'monthly_2g_8_0',
 'total_rech_num_8',
 'std_ic_t2f_mou_8',
 'std_ic_t2f_mou_7',
 'sachet_2g_6_0',
 'monthly_3g_7_0',
 'loc_ic_t2f_mou_8',
 'std_og_t2f_mou_8']

Model III¶

logr3 = sm.GLM(y_train_resampled,(sm.add_constant(X_train_resampled[selected_columns])), family = sm.families.Binomial())
logr3_fit = logr3.fit()
logr3_fit.summary()

# vif and p-values
vif(X_train_resampled, logr3_fit, selected_columns)

All features have low p-values(<0.05) and VIF (<5)
This model could be used as the interpretable logistic regression model.

Final Logistic Regression Model with RFE and Manual Elimination¶

logr3_fit.summary()

selected_columns

['sachet_3g_6_0',
 'sachet_2g_7_0',
 'sachet_2g_8_0',
 'total_rech_num_6',
 'monthly_3g_8_0',
 'monthly_2g_8_0',
 'total_rech_num_8',
 'std_ic_t2f_mou_8',
 'std_ic_t2f_mou_7',
 'sachet_2g_6_0',
 'monthly_3g_7_0',
 'loc_ic_t2f_mou_8',
 'std_og_t2f_mou_8']

# Prediction 
y_train_pred_lr = logr3_fit.predict(sm.add_constant(X_train_resampled[selected_columns]))
y_train_pred_lr.head()

0    0.118916
1    0.343873
2    0.381230
3    0.015277
4    0.001595
dtype: float64

y_test_pred_lr = logr3_fit.predict(sm.add_constant(X_test[selected_columns]))
y_test_pred_lr.head()

mobile_number
7002242818    0.013556
7000517161    0.903162
7002162382    0.247123
7002152271    0.330787
7002058655    0.056105
dtype: float64

Performance¶

Finding Optimum Probability Cutoff

# Specificity / Sensitivity Tradeoff 

# Classification at probability thresholds between 0 and 1 
y_train_pred_thres = pd.DataFrame(index=X_train_resampled.index)
thresholds = [float(x)/10 for x in range(10)]

def thresholder(x, thresh) :
    if x > thresh : 
        return 1 
    else : 
        return 0

    
for i in thresholds:
    y_train_pred_thres[i]= y_train_pred_lr.map(lambda x : thresholder(x,i))
y_train_pred_thres.head()

# DataFrame for Performance metrics at each threshold

logr_metrics_df = pd.DataFrame(columns=['sensitivity', 'specificity', 'accuracy'])
for thres,column in zip(thresholds,y_train_pred_thres.columns.to_list()) : 
    confusion = confusion_matrix(y_train_resampled, y_train_pred_thres.loc[:,column])
    sensitivity,specificity,accuracy = model_metrics_thres(confusion)
    logr_metrics_df =  logr_metrics_df.append({ 
        'sensitivity' :sensitivity,
        'specificity' : specificity,
        'accuracy' : accuracy
    }, ignore_index = True)
    
logr_metrics_df.index = thresholds
logr_metrics_df

logr_metrics_df.plot(kind='line', figsize=(24,8), grid=True, xticks=np.arange(0,1,0.02),
                title='Specificity-Sensitivity TradeOff');

The optimum probability cutoff for Logistic regression model is 0.53

optimum_cutoff = 0.53
y_train_pred_lr_final = y_train_pred_lr.map(lambda x : 1 if x > optimum_cutoff else 0)
y_test_pred_lr_final = y_test_pred_lr.map(lambda x : 1 if x > optimum_cutoff else 0)

train_matrix = confusion_matrix(y_train_resampled, y_train_pred_lr_final)
print('Confusion Matrix for train:\n', train_matrix)
test_matrix = confusion_matrix(y_test, y_test_pred_lr_final)
print('\nConfusion Matrix for test: \n', test_matrix)

Confusion Matrix for train:
 [[14531  4656]
 [ 4411 14776]]

Confusion Matrix for test: 
 [[6313 1918]
 [ 191  582]]

print('Train Performance: \n')
model_metrics(train_matrix)

print('\n\nTest Performance : \n')
model_metrics(test_matrix)

Train Performance: 

Accuracy : 0.764
Sensitivity / True Positive Rate / Recall : 0.77
Specificity / True Negative Rate :  0.757
Precision / Positive Predictive Value : 0.76
F1-score : 0.765


Test Performance : 

Accuracy : 0.766
Sensitivity / True Positive Rate / Recall : 0.753
Specificity / True Negative Rate :  0.767
Precision / Positive Predictive Value : 0.233
F1-score : 0.356

# ROC_AUC score 
print('ROC AUC score for Train : ',round(roc_auc_score(y_train_resampled, y_train_pred_lr),3), '\n' )
print('ROC AUC score for Test : ',round(roc_auc_score(y_test, y_test_pred_lr),3) )

ROC AUC score for Train :  0.843 

ROC AUC score for Test :  0.828

Model 1 : Logistic Regression (Interpretable Model Summary)¶

lr_summary_html = logr3_fit.summary().tables[1].as_html()
lr_results = pd.read_html(lr_summary_html, header=0, index_col=0)[0]
coef_column = lr_results.columns[0]
print('Most important predictors of Churn , in order of importance and their coefficients are as follows : \n')
lr_results.sort_values(by=coef_column, key=lambda x: abs(x), ascending=False)['coef']

Most important predictors of Churn , in order of importance and their coefficients are as follows :

loc_ic_t2f_mou_8   -1.2736
total_rech_num_8   -1.2033
total_rech_num_6    0.6053
monthly_3g_8_0      0.3994
monthly_2g_8_0      0.3666
std_ic_t2f_mou_8   -0.3363
std_og_t2f_mou_8   -0.2474
const              -0.2336
monthly_3g_7_0     -0.2099
std_ic_t2f_mou_7    0.1532
sachet_2g_6_0      -0.1108
sachet_2g_7_0      -0.0987
sachet_2g_8_0       0.0488
sachet_3g_6_0      -0.0399
Name: coef, dtype: float64

The above model could be used as the interpretable model for predicting telecom churn.

PCA¶

from sklearn.decomposition import PCA 
pca = PCA(random_state = 42) 
pca.fit(X_train) # note that pca is fit on original train set instead of resampled train set. 
pca.components_

array([[ 1.64887430e-01,  1.93987506e-01,  1.67239205e-01, ...,
         1.43967238e-06, -1.55704675e-06, -1.88892194e-06],
       [ 6.48591961e-02,  9.55966684e-02,  1.20775174e-01, ...,
        -2.12841595e-06, -1.47944145e-06, -3.90881587e-07],
       [ 2.38415388e-01,  2.73645507e-01,  2.38436263e-01, ...,
        -1.25598531e-06, -4.37900299e-07,  6.19889336e-07],
       ...,
       [ 1.68015588e-06,  1.93600851e-06, -1.82065762e-06, ...,
         4.25473944e-03,  2.56738368e-03,  3.51118176e-03],
       [ 0.00000000e+00, -1.11533905e-16,  1.57807487e-16, ...,
         1.73764144e-15,  6.22907679e-16,  1.45339158e-16],
       [ 0.00000000e+00,  4.98537742e-16, -6.02718139e-16, ...,
         1.27514583e-15,  1.25772226e-15,  3.41773342e-16]])

pca.explained_variance_ratio_

array([2.72067612e-01, 1.62438240e-01, 1.20827535e-01, 1.06070063e-01,
       9.11349433e-02, 4.77504400e-02, 2.63978655e-02, 2.56843982e-02,
       1.91789343e-02, 1.68045932e-02, 1.55523468e-02, 1.31676589e-02,
       1.04552128e-02, 7.72970448e-03, 7.22746863e-03, 6.14494838e-03,
       5.62073089e-03, 5.44579273e-03, 4.59009989e-03, 4.38488162e-03,
       3.46703626e-03, 3.27941490e-03, 2.78099200e-03, 2.13444270e-03,
       2.07542043e-03, 1.89794720e-03, 1.41383936e-03, 1.30240760e-03,
       1.15369576e-03, 1.05262500e-03, 9.64293417e-04, 9.16686049e-04,
       8.84067044e-04, 7.62966236e-04, 6.61794767e-04, 5.69667265e-04,
       5.12585166e-04, 5.04441248e-04, 4.82396680e-04, 4.46889495e-04,
       4.36441254e-04, 4.10389488e-04, 3.51844810e-04, 3.12626195e-04,
       2.51673027e-04, 2.34723896e-04, 1.96950034e-04, 1.71296745e-04,
       1.59882693e-04, 1.48330353e-04, 1.45919483e-04, 1.08583729e-04,
       1.04038518e-04, 8.90621848e-05, 8.53009223e-05, 7.60704088e-05,
       7.57150133e-05, 6.16615717e-05, 6.07777411e-05, 5.70517541e-05,
       5.36161089e-05, 5.28495367e-05, 5.14887086e-05, 4.73768570e-05,
       4.71283394e-05, 4.11523975e-05, 4.10392906e-05, 2.86090257e-05,
       2.19793282e-05, 1.58203581e-05, 1.50969788e-05, 1.42865579e-05,
       1.34537530e-05, 1.33026062e-05, 1.10239870e-05, 8.27539516e-06,
       7.55845974e-06, 6.45372276e-06, 6.22570067e-06, 3.42288900e-06,
       3.20804681e-06, 3.09270863e-06, 2.86608967e-06, 2.44898003e-06,
       2.08230568e-06, 1.85144734e-06, 1.64714248e-06, 1.45630245e-06,
       1.35265729e-06, 1.05472047e-06, 9.89133015e-07, 8.65864423e-07,
       7.45065121e-07, 3.66727807e-07, 6.49277820e-08, 6.13357428e-08,
       4.35995018e-08, 2.28152900e-08, 2.00441141e-08, 1.84235145e-08,
       1.66102335e-08, 1.47870989e-08, 1.23390691e-08, 1.12094165e-08,
       1.09702422e-08, 9.51924270e-09, 8.61596309e-09, 7.38051070e-09,
       7.15370081e-09, 6.29095319e-09, 5.00739371e-09, 4.68791660e-09,
       4.23376173e-09, 4.04558169e-09, 3.75847771e-09, 3.71213838e-09,
       3.32806929e-09, 3.23527525e-09, 3.12734302e-09, 2.82062311e-09,
       2.72602311e-09, 2.66103741e-09, 2.46562734e-09, 2.20243536e-09,
       2.15044476e-09, 1.59498492e-09, 1.47087974e-09, 1.06159357e-09,
       9.33938436e-10, 8.10080735e-10, 8.04656028e-10, 6.12994365e-10,
       4.82074297e-10, 4.02577318e-10, 3.58059984e-10, 3.28374076e-10,
       3.03687605e-10, 7.12091816e-11, 6.13978255e-11, 1.04375208e-33,
       1.04375208e-33])

Scree Plot¶

var_cum = np.cumsum(pca.explained_variance_ratio_)
plt.figure(figsize=(20,8))
sns.set_style('darkgrid')
sns.lineplot(np.arange(1,len(var_cum) + 1), var_cum)
plt.xticks(np.arange(0,140,5))
plt.axhline(0.95,color='r')
plt.axhline(1.0,color='r')
plt.axvline(15,color='b')
plt.axvline(45,color='b')
plt.text(10,0.96,'0.95')

plt.title('Scree Plot of Telecom Churn Train Set');

From the above scree plot, it is clear that 95% of variance in the train set can be explained by first 16 principal components and 100% of variance is explained by the first 45 principal components.

# Perform PCA using the first 45 components
pca_final = PCA(n_components=45, random_state=42)
transformed_data = pca_final.fit_transform(X_train)
X_train_pca = pd.DataFrame(transformed_data, columns=["PC_"+str(x) for x in range(1,46)], index = X_train.index)
data_train_pca = pd.concat([X_train_pca, y_train], axis=1)

data_train_pca.head()

## Plotting principal components 
sns.pairplot(data=data_train_pca, x_vars=["PC_1"], y_vars=["PC_2"], hue = "Churn", size=8);

Model 2 : PCA + Logistic Regression Model¶

# X,y Split
y_train_pca = data_train_pca.pop('Churn')
X_train_pca = data_train_pca

# Transforming test set with pca ( 45 components)
X_test_pca = pca_final.transform(X_test)

# Logistic Regression
lr_pca = LogisticRegression(random_state=100, class_weight='balanced')
lr_pca.fit(X_train_pca,y_train_pca )

LogisticRegression(class_weight='balanced', random_state=100)

# y_train predictions
y_train_pred_lr_pca = lr_pca.predict(X_train_pca)
y_train_pred_lr_pca[:5]

array([1, 0, 0, 0, 0])

# Test Prediction
X_test_pca = pca_final.transform(X_test)
y_test_pred_lr_pca = lr_pca.predict(X_test_pca)
y_test_pred_lr_pca[:5]

array([1, 1, 1, 1, 1])

Baseline Performance

train_matrix = confusion_matrix(y_train, y_train_pred_lr_pca)
test_matrix = confusion_matrix(y_test, y_test_pred_lr_pca)

print('Train Performance :\n')
model_metrics(train_matrix)

print('\nTest Performance :\n')
model_metrics(test_matrix)

Train Performance :

Accuracy : 0.645
Sensitivity / True Positive Rate / Recall : 0.905
Specificity / True Negative Rate :  0.62
Precision / Positive Predictive Value : 0.184
F1-score : 0.306

Test Performance :

Accuracy : 0.086
Sensitivity / True Positive Rate / Recall : 1.0
Specificity / True Negative Rate :  0.0
Precision / Positive Predictive Value : 0.086
F1-score : 0.158

Hyperparameter Tuning¶

# Creating a Logistic regression model using pca transformed train set
from sklearn.pipeline import Pipeline
lr_pca = LogisticRegression(random_state=100, class_weight='balanced')

from sklearn.model_selection import RandomizedSearchCV, GridSearchCV , StratifiedKFold
params = {
    'penalty' : ['l1','l2','none'], 
    'C' : [0,1,2,3,4,5,10,50]
}
folds = StratifiedKFold(n_splits=4, shuffle=True, random_state=100)

search = GridSearchCV(cv=folds, estimator = lr_pca, param_grid=params,scoring='roc_auc', verbose=True, n_jobs=-1)
search.fit(X_train_pca, y_train_pca)

Fitting 4 folds for each of 24 candidates, totalling 96 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.
[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:    4.0s
[Parallel(n_jobs=-1)]: Done  96 out of  96 | elapsed:    6.9s finished

GridSearchCV(cv=StratifiedKFold(n_splits=4, random_state=100, shuffle=True),
             estimator=LogisticRegression(class_weight='balanced',
                                          random_state=100),
             n_jobs=-1,
             param_grid={'C': [0, 1, 2, 3, 4, 5, 10, 50],
                         'penalty': ['l1', 'l2', 'none']},
             scoring='roc_auc', verbose=True)

# Optimum Hyperparameters
print('Best ROC-AUC score :', search.best_score_)
print('Best Parameters :', search.best_params_)

Best ROC-AUC score : 0.8763924253372933
Best Parameters : {'C': 0, 'penalty': 'none'}

# Modelling using the best LR-PCA estimator 
lr_pca_best = search.best_estimator_
lr_pca_best_fit = lr_pca_best.fit(X_train_pca, y_train_pca)

# Prediction on Train set
y_train_pred_lr_pca_best = lr_pca_best_fit.predict(X_train_pca)
y_train_pred_lr_pca_best[:5]

array([1, 1, 0, 0, 0])

# Prediction on test set
y_test_pred_lr_pca_best = lr_pca_best_fit.predict(X_test_pca)
y_test_pred_lr_pca_best[:5]

array([1, 1, 1, 1, 1])

## Model Performance after Hyper Parameter Tuning

train_matrix = confusion_matrix(y_train, y_train_pred_lr_pca_best)
test_matrix = confusion_matrix(y_test, y_test_pred_lr_pca_best)

print('Train Performance :\n')
model_metrics(train_matrix)

print('\nTest Performance :\n')
model_metrics(test_matrix)

Train Performance :

Accuracy : 0.627
Sensitivity / True Positive Rate / Recall : 0.918
Specificity / True Negative Rate :  0.599
Precision / Positive Predictive Value : 0.179
F1-score : 0.3

Test Performance :

Accuracy : 0.086
Sensitivity / True Positive Rate / Recall : 1.0
Specificity / True Negative Rate :  0.0
Precision / Positive Predictive Value : 0.086
F1-score : 0.158

Model 3 : PCA + Random Forest¶

from sklearn.ensemble import RandomForestClassifier

# creating a random forest classifier using pca output

pca_rf = RandomForestClassifier(random_state=42, class_weight= {0 : class_1/(class_0 + class_1) , 1 : class_0/(class_0 + class_1) } , oob_score=True, n_jobs=-1,verbose=1)
pca_rf

RandomForestClassifier(class_weight={0: 0.08640165272733331,
                                     1: 0.9135983472726666},
                       n_jobs=-1, oob_score=True, random_state=42, verbose=1)

# Hyper parameter Tuning
params = {
    'n_estimators'  : [30,40,50,100],
    'max_depth' : [3,4,5,6,7],
    'min_samples_leaf' : [15,20,25,30]
}
folds = StratifiedKFold(n_splits=4, shuffle=True, random_state=42)
pca_rf_model_search = GridSearchCV(estimator=pca_rf, param_grid=params, 
                                   cv=folds, scoring='roc_auc', verbose=True, n_jobs=-1 )

pca_rf_model_search.fit(X_train_pca, y_train)

Fitting 4 folds for each of 80 candidates, totalling 320 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.
[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:   23.2s
[Parallel(n_jobs=-1)]: Done 192 tasks      | elapsed:  2.7min
[Parallel(n_jobs=-1)]: Done 320 out of 320 | elapsed:  5.5min finished
[Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 4 concurrent workers.
[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:    1.2s
[Parallel(n_jobs=-1)]: Done 100 out of 100 | elapsed:    2.6s finished

GridSearchCV(cv=StratifiedKFold(n_splits=4, random_state=42, shuffle=True),
             estimator=RandomForestClassifier(class_weight={0: 0.08640165272733331,
                                                            1: 0.9135983472726666},
                                              n_jobs=-1, oob_score=True,
                                              random_state=42, verbose=1),
             n_jobs=-1,
             param_grid={'max_depth': [3, 4, 5, 6, 7],
                         'min_samples_leaf': [15, 20, 25, 30],
                         'n_estimators': [30, 40, 50, 100]},
             scoring='roc_auc', verbose=True)

# Optimum Hyperparameters
print('Best ROC-AUC score :', pca_rf_model_search.best_score_)
print('Best Parameters :', pca_rf_model_search.best_params_)

Best ROC-AUC score : 0.8861621751601011
Best Parameters : {'max_depth': 7, 'min_samples_leaf': 20, 'n_estimators': 100}

# Modelling using the best PCA-RandomForest Estimator 
pca_rf_best = pca_rf_model_search.best_estimator_
pca_rf_best_fit = pca_rf_best.fit(X_train_pca, y_train)

# Prediction on Train set
y_train_pred_pca_rf_best = pca_rf_best_fit.predict(X_train_pca)
y_train_pred_pca_rf_best[:5]

[Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 4 concurrent workers.
[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:    1.1s
[Parallel(n_jobs=-1)]: Done 100 out of 100 | elapsed:    2.7s finished
[Parallel(n_jobs=4)]: Using backend ThreadingBackend with 4 concurrent workers.
[Parallel(n_jobs=4)]: Done  42 tasks      | elapsed:    0.0s
[Parallel(n_jobs=4)]: Done 100 out of 100 | elapsed:    0.1s finished

array([0, 0, 0, 0, 0])

# Prediction on test set
y_test_pred_pca_rf_best = pca_rf_best_fit.predict(X_test_pca)
y_test_pred_pca_rf_best[:5]

[Parallel(n_jobs=4)]: Using backend ThreadingBackend with 4 concurrent workers.
[Parallel(n_jobs=4)]: Done  42 tasks      | elapsed:    0.1s
[Parallel(n_jobs=4)]: Done 100 out of 100 | elapsed:    0.1s finished

array([0, 0, 0, 0, 0])

## PCA - RandomForest Model Performance - Hyper Parameter Tuned

train_matrix = confusion_matrix(y_train, y_train_pred_pca_rf_best)
test_matrix = confusion_matrix(y_test, y_test_pred_pca_rf_best)

print('Train Performance :\n')
model_metrics(train_matrix)

print('\nTest Performance :\n')
model_metrics(test_matrix)

Train Performance :

Accuracy : 0.882
Sensitivity / True Positive Rate / Recall : 0.816
Specificity / True Negative Rate :  0.888
Precision / Positive Predictive Value : 0.408
F1-score : 0.544

Test Performance :

Accuracy : 0.86
Sensitivity / True Positive Rate / Recall : 0.80
Specificity / True Negative Rate :  0.78
Precision / Positive Predictive Value : 0.37
F1-score : 0.51

## out of bag error 
pca_rf_best_fit.oob_score_

0.8625220164707003

Model 4 : PCA + XGBoost¶

import xgboost as xgb
pca_xgb = xgb.XGBClassifier(random_state=42, scale_pos_weight= class_0/class_1 ,
                                    tree_method='hist', 
                                   objective='binary:logistic',
                                  
                                  
                                  )  # scale_pos_weight takes care of class imbalance
pca_xgb.fit(X_train_pca, y_train)

XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,
              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,
              importance_type='gain', interaction_constraints='',
              learning_rate=0.300000012, max_delta_step=0, max_depth=6,
              min_child_weight=1, missing=nan, monotone_constraints='()',
              n_estimators=100, n_jobs=0, num_parallel_tree=1, random_state=42,
              reg_alpha=0, reg_lambda=1, scale_pos_weight=10.573852680293097,
              subsample=1, tree_method='hist', validate_parameters=1,
              verbosity=None)

print('Baseline Train AUC Score')
roc_auc_score(y_train, pca_xgb.predict_proba(X_train_pca)[:, 1])

Baseline Train AUC Score

0.9999996277241286

print('Baseline Test AUC Score')
roc_auc_score(y_test, pca_xgb.predict_proba(X_test_pca)[:, 1])

Baseline Test AUC Score

0.46093390352284136

## Hyper parameter Tuning
parameters = {
              'learning_rate': [0.1, 0.2, 0.3],
              'gamma' : [10,20,50],
              'max_depth': [2,3,4],
              'min_child_weight': [25,50],
              'n_estimators': [150,200,500]}
pca_xgb_search = GridSearchCV(estimator=pca_xgb , param_grid=parameters,scoring='roc_auc', cv=folds, n_jobs=-1, verbose=1)
pca_xgb_search.fit(X_train_pca, y_train)

Fitting 4 folds for each of 162 candidates, totalling 648 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.
[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:   28.3s
[Parallel(n_jobs=-1)]: Done 192 tasks      | elapsed:  2.1min
[Parallel(n_jobs=-1)]: Done 442 tasks      | elapsed:  4.8min
[Parallel(n_jobs=-1)]: Done 648 out of 648 | elapsed:  8.0min finished

GridSearchCV(cv=StratifiedKFold(n_splits=4, random_state=42, shuffle=True),
             estimator=XGBClassifier(base_score=0.5, booster='gbtree',
                                     colsample_bylevel=1, colsample_bynode=1,
                                     colsample_bytree=1, gamma=0, gpu_id=-1,
                                     importance_type='gain',
                                     interaction_constraints='',
                                     learning_rate=0.300000012,
                                     max_delta_step=0, max_depth=6,
                                     min_child_weight=1, missing=nan,
                                     monotone_...
                                     n_estimators=100, n_jobs=0,
                                     num_parallel_tree=1, random_state=42,
                                     reg_alpha=0, reg_lambda=1,
                                     scale_pos_weight=10.573852680293097,
                                     subsample=1, tree_method='hist',
                                     validate_parameters=1, verbosity=None),
             n_jobs=-1,
             param_grid={'gamma': [10, 20, 50],
                         'learning_rate': [0.1, 0.2, 0.3],
                         'max_depth': [2, 3, 4], 'min_child_weight': [25, 50],
                         'n_estimators': [150, 200, 500]},
             scoring='roc_auc', verbose=1)

# Optimum Hyperparameters
print('Best ROC-AUC score :', pca_xgb_search.best_score_)
print('Best Parameters :', pca_xgb_search.best_params_)

Best ROC-AUC score : 0.8955777259491308
Best Parameters : {'gamma': 10, 'learning_rate': 0.1, 'max_depth': 2, 'min_child_weight': 50, 'n_estimators': 500}

# Modelling using the best PCA-XGBoost Estimator 
pca_xgb_best = pca_xgb_search.best_estimator_
pca_xgb_best_fit = pca_xgb_best.fit(X_train_pca, y_train)

# Prediction on Train set
y_train_pred_pca_xgb_best = pca_xgb_best_fit.predict(X_train_pca)
y_train_pred_pca_xgb_best[:5]

array([0, 0, 0, 0, 0])

X_train_pca.head()

# Prediction on test set
X_test_pca = pca_final.transform(X_test)
X_test_pca = pd.DataFrame(X_test_pca, index=X_test.index, columns = X_train_pca.columns)
y_test_pred_pca_xgb_best = pca_xgb_best_fit.predict(X_test_pca)
y_test_pred_pca_xgb_best[:5]

array([1, 1, 1, 1, 1])

## PCA - XGBOOST [Hyper parameter tuned] Model Performance

train_matrix = confusion_matrix(y_train, y_train_pred_pca_xgb_best)
test_matrix = confusion_matrix(y_test, y_test_pred_pca_xgb_best)

print('Train Performance :\n')
model_metrics(train_matrix)

print('\nTest Performance :\n')
model_metrics(test_matrix)

Train Performance :

Accuracy : 0.873
Sensitivity / True Positive Rate / Recall : 0.887
Specificity / True Negative Rate :  0.872
Precision / Positive Predictive Value : 0.396
F1-score : 0.548

Test Performance :

Accuracy : 0.086
Sensitivity / True Positive Rate / Recall : 1.0
Specificity / True Negative Rate :  0.0
Precision / Positive Predictive Value : 0.086
F1-score : 0.158

## PCA - XGBOOST [Hyper parameter tuned] Model Performance
print('Train AUC Score')
print(roc_auc_score(y_train, pca_xgb_best.predict_proba(X_train_pca)[:, 1]))
print('Test AUC Score')
print(roc_auc_score(y_test, pca_xgb_best.predict_proba(X_test_pca)[:, 1]))

Train AUC Score
0.9442462043611259
Test AUC Score
0.6353301334697982

Recommendations¶

print('Most Important Predictors of churn , in the order of importance are : ')
lr_results.sort_values(by=coef_column, key=lambda x: abs(x), ascending=False)['coef']

Most Important Predictors of churn , in the order of importance are :

loc_ic_t2f_mou_8   -1.2736
total_rech_num_8   -1.2033
total_rech_num_6    0.6053
monthly_3g_8_0      0.3994
monthly_2g_8_0      0.3666
std_ic_t2f_mou_8   -0.3363
std_og_t2f_mou_8   -0.2474
const              -0.2336
monthly_3g_7_0     -0.2099
std_ic_t2f_mou_7    0.1532
sachet_2g_6_0      -0.1108
sachet_2g_7_0      -0.0987
sachet_2g_8_0       0.0488
sachet_3g_6_0      -0.0399
Name: coef, dtype: float64

From the above, the following are the strongest indicators of churn

Customers who churn show lower average monthly local incoming calls from fixed line in the action period by 1.27 standard deviations , compared to users who don't churn , when all other factors are held constant. This is the strongest indicator of churn.
Customers who churn show lower number of recharges done in action period by 1.20 standard deviations, when all other factors are held constant. This is the second strongest indicator of churn.
Further customers who churn have done 0.6 standard deviations higher recharge than non-churn customers. This factor when coupled with above factors is a good indicator of churn.
Customers who churn are more likely to be users of 'monthly 2g package-0 / monthly 3g package-0' in action period (approximately 0.3 std deviations higher than other packages), when all other factors are held constant.

Based on the above indicators the recommendations to the telecom company are :

Concentrate on users with 1.27 std devations lower than average incoming calls from fixed line. They are most likely to churn.
Concentrate on users who recharge less number of times ( less than 1.2 std deviations compared to avg) in the 8th month. They are second most likely to churn.
Models with high sensitivity are the best for predicting churn. Use the PCA + Logistic Regression model to predict churn. It has an ROC score of 0.87, test sensitivity of 100%

	date_of_last_rech_6	date_of_last_rech_7	date_of_last_rech_8	date_of_last_rech_9	date_of_last_rech_data_6	date_of_last_rech_data_7	date_of_last_rech_data_8	date_of_last_rech_data_9
mobile_number
7000842753	6/21/2014	7/16/2014	8/8/2014	9/28/2014	6/21/2014	7/16/2014	8/8/2014	NaN
7001865778	6/29/2014	7/31/2014	8/28/2014	9/30/2014	NaN	7/25/2014	8/10/2014	NaN
7001625959	6/17/2014	7/24/2014	8/14/2014	9/29/2014	NaN	NaN	NaN	9/17/2014
7001204172	6/28/2014	7/31/2014	8/31/2014	9/30/2014	NaN	NaN	NaN	NaN
7000142493	6/26/2014	7/28/2014	8/9/2014	9/28/2014	6/4/2014	NaN	NaN	NaN

	date_of_last_rech_6	date_of_last_rech_7	date_of_last_rech_8	date_of_last_rech_9	date_of_last_rech_data_6	date_of_last_rech_data_7	date_of_last_rech_data_8	date_of_last_rech_data_9
mobile_number
7000842753	2014-06-21	2014-07-16	2014-08-08	2014-09-28	2014-06-21	2014-07-16	2014-08-08	NaT
7001865778	2014-06-29	2014-07-31	2014-08-28	2014-09-30	NaT	2014-07-25	2014-08-10	NaT
7001625959	2014-06-17	2014-07-24	2014-08-14	2014-09-29	NaT	NaT	NaT	2014-09-17
7001204172	2014-06-28	2014-07-31	2014-08-31	2014-09-30	NaT	NaT	NaT	NaT
7000142493	2014-06-26	2014-07-28	2014-08-09	2014-09-28	2014-06-04	NaT	NaT	NaT

	Datatype	Non_Null_Count	Null_Count	Null_Percentage	Unique_Values_Count
av_rech_amt_data_6	float64	11397	18614	62.02	794
count_rech_3g_6	category	11397	18614	62.02	25
count_rech_2g_6	float64	11397	18614	62.02	30
arpu_2g_6	float64	11397	18614	62.02	4503
max_rech_data_6	float64	11397	18614	62.02	43
night_pck_user_6	category	11397	18614	62.02	2
date_of_last_rech_data_6	datetime64[ns]	11397	18614	62.02	30
total_rech_data_6	float64	11397	18614	62.02	36
arpu_3g_6	float64	11397	18614	62.02	4875
fb_user_6	category	11397	18614	62.02	2
max_rech_data_9	float64	11461	18550	61.81	48
count_rech_3g_9	category	11461	18550	61.81	27
fb_user_9	category	11461	18550	61.81	2
total_rech_data_9	float64	11461	18550	61.81	35
date_of_last_rech_data_9	datetime64[ns]	11461	18550	61.81	30
av_rech_amt_data_9	float64	11461	18550	61.81	812
arpu_2g_9	float64	11461	18550	61.81	3846
arpu_3g_9	float64	11461	18550	61.81	4800
night_pck_user_9	category	11461	18550	61.81	2
count_rech_2g_9	float64	11461	18550	61.81	29
fb_user_7	category	11662	18349	61.14	2
date_of_last_rech_data_7	datetime64[ns]	11662	18349	61.14	31
total_rech_data_7	float64	11662	18349	61.14	40
night_pck_user_7	category	11662	18349	61.14	2
max_rech_data_7	float64	11662	18349	61.14	46
count_rech_2g_7	float64	11662	18349	61.14	35
arpu_3g_7	float64	11662	18349	61.14	4860
av_rech_amt_data_7	float64	11662	18349	61.14	863
arpu_2g_7	float64	11662	18349	61.14	4219
count_rech_3g_7	category	11662	18349	61.14	28
night_pck_user_8	category	11754	18257	60.83	2
fb_user_8	category	11754	18257	60.83	2
arpu_2g_8	float64	11754	18257	60.83	3854
count_rech_2g_8	float64	11754	18257	60.83	33
date_of_last_rech_data_8	datetime64[ns]	11754	18257	60.83	31
av_rech_amt_data_8	float64	11754	18257	60.83	837
arpu_3g_8	float64	11754	18257	60.83	4769
total_rech_data_8	float64	11754	18257	60.83	45
count_rech_3g_8	category	11754	18257	60.83	29
max_rech_data_8	float64	11754	18257	60.83	47

	arpu_6	arpu_7	arpu_8	onnet_mou_6	onnet_mou_7	onnet_mou_8	offnet_mou_6	offnet_mou_7	offnet_mou_8	roam_ic_mou_6	roam_ic_mou_7	roam_ic_mou_8	roam_og_mou_6	roam_og_mou_7	roam_og_mou_8	loc_og_t2t_mou_6	loc_og_t2t_mou_7	loc_og_t2t_mou_8	loc_og_t2m_mou_6	loc_og_t2m_mou_7	loc_og_t2m_mou_8	loc_og_t2f_mou_6	loc_og_t2f_mou_7	loc_og_t2f_mou_8	loc_og_t2c_mou_6	loc_og_t2c_mou_7	loc_og_t2c_mou_8	loc_og_mou_6	loc_og_mou_7	loc_og_mou_8	std_og_t2t_mou_6	std_og_t2t_mou_7	std_og_t2t_mou_8	std_og_t2m_mou_6	std_og_t2m_mou_7	std_og_t2m_mou_8	std_og_t2f_mou_6	std_og_t2f_mou_7	std_og_t2f_mou_8	std_og_mou_6	std_og_mou_7	std_og_mou_8	isd_og_mou_6	isd_og_mou_7	isd_og_mou_8	spl_og_mou_6	spl_og_mou_7	spl_og_mou_8	og_others_6	og_others_7	og_others_8	total_og_mou_6	total_og_mou_7	total_og_mou_8	loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	loc_ic_t2t_mou_8	loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	loc_ic_mou_6	loc_ic_mou_7	loc_ic_mou_8	std_ic_t2t_mou_6	std_ic_t2t_mou_7	std_ic_t2t_mou_8	std_ic_t2m_mou_6	std_ic_t2m_mou_7	std_ic_t2m_mou_8	std_ic_t2f_mou_6	std_ic_t2f_mou_7	std_ic_t2f_mou_8	std_ic_mou_6	std_ic_mou_7	std_ic_mou_8	total_ic_mou_6	total_ic_mou_7	total_ic_mou_8	spl_ic_mou_6	spl_ic_mou_7	spl_ic_mou_8	isd_ic_mou_6	isd_ic_mou_7	isd_ic_mou_8	ic_others_6	ic_others_7	ic_others_8	total_rech_num_6	total_rech_num_7	total_rech_num_8	total_rech_amt_6	total_rech_amt_7	total_rech_amt_8	max_rech_amt_6	max_rech_amt_7	max_rech_amt_8	last_day_rch_amt_6	last_day_rch_amt_7	last_day_rch_amt_8	vol_2g_mb_6	vol_2g_mb_7	vol_2g_mb_8	vol_3g_mb_6	vol_3g_mb_7	vol_3g_mb_8	aon	vbc_3g_8	vbc_3g_7	vbc_3g_6	Average_rech_amt_6n7
count	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.00000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.00000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.00000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000	30011.000000
mean	587.284404	589.135427	534.857433	296.034461	304.343206	267.600412	417.933372	423.924375	375.021691	17.412764	13.522114	13.25627	29.321648	22.036003	21.469272	94.680696	95.729729	87.139995	181.279583	181.271524	167.591199	6.97933	7.097268	6.494314	1.567160	1.862229	1.712739	282.948414	284.107492	261.233938	189.753131	199.877508	172.196408	203.097767	213.411914	179.568790	2.010766	2.034241	1.789728	394.865994	415.327988	353.558826	2.264425	2.207400	2.029314	5.916364	7.425487	6.885193	0.692507	0.047600	0.059131	686.697541	709.124730	623.774684	68.749054	70.311351	65.936968	159.613810	160.813032	153.628517	15.595629	16.510023	14.706512	243.968340	247.644401	234.281577	16.229350	16.893723	15.051559	32.015163	33.477150	30.434765	2.874506	2.992948	2.680925	51.122992	53.36786	48.170990	307.512073	314.875472	295.426531	0.066731	0.018066	0.027660	11.156530	12.360190	11.700835	1.188803	1.476889	1.237756	12.121322	11.913465	10.225317	697.365833	695.962880	613.638799	171.414048	175.661058	162.869348	104.485655	105.287128	95.653294	78.859009	78.171382	69.209105	258.392681	278.093737	269.864111	1264.064776	129.439626	135.127102	121.360548	696.664356
std	442.722413	462.897814	492.259586	460.775592	481.780488	466.560947	470.588583	486.525332	477.489377	79.152657	76.303736	74.55207	118.570414	97.925249	106.244774	236.849265	248.132623	234.721938	250.132066	240.722132	234.862468	22.66552	22.588864	20.220028	6.889317	9.255645	7.397562	379.985249	375.837282	366.539171	409.716719	428.119476	410.033964	413.489240	437.941904	416.752834	12.457422	13.350441	11.700376	606.508681	637.446710	616.219690	45.918087	45.619381	44.794926	18.621373	23.065743	22.893414	2.281325	2.741786	3.320320	660.356820	685.071178	685.983313	158.647160	167.315954	155.702334	222.001036	219.432004	217.026349	45.827009	49.478371	43.714061	312.805586	315.468343	307.043800	78.862358	84.691403	72.433104	101.084965	105.806605	105.308898	19.928472	20.511317	20.269535	140.504104	149.17944	140.965196	361.159561	369.654489	360.343153	0.194273	0.181944	0.116574	67.258387	76.992293	74.928607	13.987003	15.406483	12.889879	9.543550	9.605532	9.478572	539.325984	562.143146	601.821630	174.703215	181.545389	172.605809	142.767207	141.148386	145.260363	277.445058	280.331857	268.494284	866.195376	855.682340	859.299266	975.263117	390.478591	408.024394	389.726031	488.782088
min	-2258.709000	-2014.045000	-945.808000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.00000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.00000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.00000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	180.000000	0.000000	0.000000	0.000000	368.500000
25%	364.161000	365.004500	289.609500	41.110000	40.950000	27.010000	137.335000	135.680000	95.695000	0.000000	0.000000	0.00000	0.000000	0.000000	0.000000	8.320000	9.130000	5.790000	30.290000	33.580000	22.420000	0.00000	0.000000	0.000000	0.000000	0.000000	0.000000	51.010000	56.710000	38.270000	0.000000	0.000000	0.000000	1.600000	1.330000	0.000000	0.000000	0.000000	0.000000	5.950000	5.555000	1.780000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	266.170000	275.045000	188.790000	8.290000	9.460000	6.810000	33.460000	38.130000	29.660000	0.000000	0.000000	0.000000	56.700000	63.535000	49.985000	0.000000	0.000000	0.000000	0.450000	0.480000	0.000000	0.000000	0.000000	0.000000	2.630000	2.78000	1.430000	89.975000	98.820000	78.930000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	6.000000	6.000000	4.000000	432.000000	426.500000	309.000000	110.000000	110.000000	67.000000	30.000000	27.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	480.000000	0.000000	0.000000	0.000000	450.000000
50%	495.682000	493.561000	452.091000	125.830000	125.460000	99.440000	282.190000	281.940000	240.940000	0.000000	0.000000	0.00000	0.000000	0.000000	0.000000	32.590000	33.160000	28.640000	101.240000	104.340000	89.810000	0.33000	0.400000	0.160000	0.000000	0.000000	0.000000	166.310000	170.440000	148.280000	12.830000	13.350000	5.930000	37.730000	37.530000	23.660000	0.000000	0.000000	0.000000	126.010000	131.730000	72.890000	0.000000	0.000000	0.000000	0.210000	0.780000	0.490000	0.000000	0.000000	0.000000	510.230000	525.580000	435.330000	29.130000	30.130000	26.840000	93.940000	96.830000	89.810000	1.960000	2.210000	1.850000	151.060000	154.830000	142.840000	1.050000	1.200000	0.560000	7.080000	7.460000	5.710000	0.000000	0.000000	0.000000	15.030000	16.11000	12.560000	205.240000	211.190000	193.440000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	9.000000	9.000000	8.000000	584.000000	581.000000	520.000000	120.000000	128.000000	130.000000	110.000000	98.000000	50.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	914.000000	0.000000	0.000000	0.000000	568.500000
75%	703.922000	700.788000	671.150000	353.310000	359.925000	297.735000	523.125000	532.695000	482.610000	0.000000	0.000000	0.00000	0.000000	0.000000	0.000000	91.460000	91.480000	84.670000	240.165000	239.485000	223.590000	5.09000	5.260000	4.680000	0.000000	0.100000	0.050000	374.475000	375.780000	348.310000	178.085000	191.380000	132.820000	211.210000	223.010000	164.725000	0.000000	0.000000	0.000000	573.090000	615.150000	481.030000	0.000000	0.000000	0.000000	5.160000	7.110000	6.380000	0.000000	0.000000	0.000000	899.505000	931.050000	833.100000	73.640000	74.680000	70.330000	202.830000	203.485000	196.975000	12.440000	13.035000	11.605000	315.500000	316.780000	302.110000	10.280000	10.980000	8.860000	27.540000	29.235000	25.330000	0.180000	0.260000	0.130000	47.540000	50.36000	43.410000	393.680000	396.820000	380.410000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.060000	0.000000	0.060000	15.000000	15.000000	13.000000	837.000000	835.000000	790.000000	200.000000	200.000000	198.000000	120.000000	130.000000	130.000000	14.450000	14.960000	9.620000	0.000000	2.080000	0.000000	1924.000000	1.600000	1.990000	0.000000	795.500000
max	27731.088000	35145.834000	33543.624000	7376.710000	8157.780000	10752.560000	8362.360000	9667.130000	14007.340000	2613.310000	3813.290000	4169.81000	3775.110000	2812.040000	5337.040000	6431.330000	7400.660000	10752.560000	4729.740000	4557.140000	4961.330000	1466.03000	1196.430000	928.490000	342.860000	569.710000	351.830000	10643.380000	7674.780000	11039.910000	7366.580000	8133.660000	8014.430000	8314.760000	9284.740000	13950.040000	628.560000	544.630000	516.910000	8432.990000	10936.730000	13980.060000	5900.660000	5490.280000	5681.540000	1023.210000	1265.790000	1390.880000	100.610000	370.130000	394.930000	10674.030000	11365.310000	14043.060000	6351.440000	5709.590000	4003.210000	4693.860000	4388.730000	5738.460000	1678.410000	1983.010000	1588.530000	6496.110000	6466.740000	5748.810000	5459.560000	5800.930000	4309.290000	4630.230000	3470.380000	5645.860000	1351.110000	1136.080000	1394.890000	5459.630000	6745.76000	5957.140000	6798.640000	7279.080000	5990.710000	19.760000	21.330000	6.230000	3965.690000	4747.910000	4100.380000	1344.140000	1495.940000	1209.860000	307.000000	138.000000	196.000000	35190.000000	40335.000000	45320.000000	4010.000000	4010.000000	4449.000000	4010.000000	4010.000000	4449.000000	10285.900000	7873.550000	11117.610000	45735.400000	28144.120000	30036.060000	4321.000000	12916.220000	9165.600000	11166.210000	37762.500000

	onnet_mou_6	onnet_mou_7	onnet_mou_8	offnet_mou_6	offnet_mou_7	offnet_mou_8	roam_ic_mou_6	roam_ic_mou_7	roam_ic_mou_8	roam_og_mou_6	roam_og_mou_7	roam_og_mou_8	loc_og_t2t_mou_6	loc_og_t2t_mou_7	loc_og_t2t_mou_8	loc_og_t2m_mou_6	loc_og_t2m_mou_7	loc_og_t2m_mou_8	loc_og_t2f_mou_6	loc_og_t2f_mou_7	loc_og_t2f_mou_8	loc_og_t2c_mou_6	loc_og_t2c_mou_7	loc_og_t2c_mou_8	loc_og_mou_6	loc_og_mou_7	loc_og_mou_8	std_og_t2t_mou_6	std_og_t2t_mou_7	std_og_t2t_mou_8	std_og_t2m_mou_6	std_og_t2m_mou_7	std_og_t2m_mou_8	std_og_t2f_mou_6	std_og_t2f_mou_7	std_og_t2f_mou_8	std_og_mou_6	std_og_mou_7	std_og_mou_8	isd_og_mou_6	isd_og_mou_7	isd_og_mou_8	spl_og_mou_6	spl_og_mou_7	spl_og_mou_8	og_others_6	og_others_7	og_others_8	loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	loc_ic_t2t_mou_8	loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	loc_ic_mou_6	loc_ic_mou_7	loc_ic_mou_8	std_ic_t2t_mou_6	std_ic_t2t_mou_7	std_ic_t2t_mou_8	std_ic_t2m_mou_6	std_ic_t2m_mou_7	std_ic_t2m_mou_8	std_ic_t2f_mou_6	std_ic_t2f_mou_7	std_ic_t2f_mou_8	std_ic_mou_6	std_ic_mou_7	std_ic_mou_8	spl_ic_mou_6	spl_ic_mou_7	spl_ic_mou_8	isd_ic_mou_6	isd_ic_mou_7	isd_ic_mou_8	ic_others_6	ic_others_7	ic_others_8	total_rech_num_6	total_rech_num_7	total_rech_num_8	max_rech_amt_6	max_rech_amt_7	max_rech_amt_8	last_day_rch_amt_6	last_day_rch_amt_7	last_day_rch_amt_8	aon	Average_rech_amt_6n7	delta_vol_2g	delta_vol_3g	delta_total_og_mou	delta_total_ic_mou	delta_vbc_3g	delta_arpu	delta_total_rech_amt
0.9	794.98	824.38	723.61	915.58	935.69	853.79	32.73	18.36	18.68	64.48	41.20	37.11	207.93	207.84	196.91	435.16	437.49	416.66	18.38	18.66	16.96	4.04	4.84	4.45	661.74	657.38	633.34	630.53	663.79	567.34	604.41	645.88	531.26	2.20	2.18	1.73	1140.93	1177.18	1057.29	0.00	0.00	0.00	15.93	19.51	18.04	2.26	0.00	0.00	154.88	156.61	148.14	368.54	364.54	360.54	39.23	41.04	37.19	559.28	558.99	549.79	34.73	36.01	32.14	73.38	75.28	68.58	4.36	4.58	3.94	115.91	118.66	108.38	0.28	0.00	0.00	15.01	18.30	15.33	1.16	1.59	1.23	23.00	23.00	21.00	297.00	300.00	252.00	250.00	250.00	225.00	2846.00	1118.00	29.84	170.07	345.07	147.30	69.83	257.31	319.00
0.91	848.97	878.35	783.49	966.74	984.02	899.29	39.69	23.28	23.39	78.43	50.01	46.44	225.96	224.87	213.83	461.10	461.81	441.84	20.28	20.68	18.84	4.68	5.51	5.11	703.11	692.67	669.63	686.26	722.84	622.13	658.47	695.77	583.42	2.91	2.80	2.28	1195.61	1244.40	1125.28	0.00	0.00	0.00	17.54	21.28	19.69	2.54	0.00	0.00	165.79	168.03	159.84	390.64	387.11	382.20	43.59	45.39	41.21	593.13	589.65	580.54	38.21	39.91	35.93	80.41	81.93	75.54	5.21	5.49	4.71	125.98	129.29	118.24	0.30	0.00	0.00	18.34	21.84	18.83	1.44	1.94	1.51	24.00	24.00	22.00	325.00	330.00	289.00	250.00	250.00	250.00	2910.10	1156.00	39.88	227.15	377.46	161.80	95.33	278.90	345.50
0.92	909.05	941.99	848.96	1031.39	1038.09	953.35	48.71	29.68	29.64	93.60	60.97	57.59	247.94	244.78	232.33	490.63	488.04	468.83	22.56	23.14	20.93	5.45	6.26	5.86	742.96	735.69	711.57	750.31	786.39	680.10	713.49	760.98	640.57	3.74	3.71	3.01	1268.83	1315.08	1201.29	0.13	0.05	0.00	19.26	23.39	21.78	2.86	0.00	0.00	180.18	181.49	173.59	415.89	412.03	405.97	48.65	50.66	46.19	629.64	624.36	614.45	42.73	44.58	39.99	88.27	90.41	83.44	6.33	6.61	5.75	138.32	142.16	130.55	0.33	0.00	0.03	22.58	26.94	23.58	1.78	2.38	1.86	25.00	25.00	23.00	350.00	350.00	330.00	250.00	250.00	250.00	2981.20	1202.00	53.66	289.30	419.97	177.35	127.50	303.51	375.00
0.93	990.48	1016.15	920.96	1094.77	1103.93	1017.35	60.42	37.28	37.90	111.15	75.00	72.45	275.51	271.70	254.64	523.56	519.80	500.38	25.25	26.00	23.51	6.34	7.15	6.79	794.01	786.73	759.45	812.08	856.34	753.44	777.69	828.18	706.23	4.89	4.70	4.00	1358.41	1404.59	1283.20	0.33	0.25	0.00	21.33	25.84	24.07	3.23	0.00	0.00	195.66	196.99	188.25	444.94	439.30	434.35	54.59	57.50	51.92	671.69	667.07	654.41	48.03	50.82	45.64	98.00	99.68	93.47	7.84	8.08	7.08	153.30	158.86	146.87	0.36	0.00	0.11	28.18	33.33	29.95	2.20	2.93	2.38	27.00	27.00	25.00	350.00	398.00	350.00	252.00	252.00	250.00	3055.30	1257.00	71.44	361.57	466.79	195.08	166.36	331.83	408.50
0.9400000000000001	1066.85	1097.12	1007.56	1168.09	1186.36	1096.62	73.96	49.44	48.29	137.40	94.73	90.27	307.64	305.87	282.50	563.70	559.41	537.10	29.13	29.75	26.89	7.36	8.45	7.84	855.97	849.42	817.15	888.23	933.58	842.48	856.37	907.95	786.16	6.23	6.11	5.26	1456.41	1503.09	1385.91	0.63	0.51	0.23	23.92	28.49	26.83	3.64	0.00	0.00	216.28	214.47	207.73	478.56	478.91	472.15	62.96	65.76	58.84	717.00	711.74	704.89	55.49	58.04	52.58	110.61	113.20	105.54	9.66	9.89	8.77	174.34	179.84	165.92	0.40	0.00	0.21	35.80	40.97	36.57	2.81	3.73	2.98	28.00	28.00	26.00	400.00	455.00	398.00	252.00	252.00	252.00	3107.00	1317.70	97.46	463.36	524.65	218.73	217.42	366.26	447.20
0.9500000000000001	1153.97	1208.17	1115.66	1271.47	1286.28	1188.46	94.59	63.34	62.80	168.46	119.34	114.80	348.62	346.90	324.14	614.99	608.01	585.06	33.59	34.09	31.31	8.69	9.95	9.33	935.51	920.12	883.25	986.24	1029.29	936.49	960.80	1004.26	886.56	8.16	7.92	7.18	1558.50	1624.81	1518.82	1.10	1.01	0.55	26.81	32.15	30.23	4.14	0.00	0.00	243.94	238.62	232.50	520.55	518.65	516.67	72.61	76.05	67.56	773.27	781.18	767.31	64.94	66.55	61.56	126.66	130.41	121.88	12.24	12.31	10.98	200.64	205.16	191.95	0.43	0.06	0.25	46.45	51.98	46.48	3.63	4.83	3.93	30.00	30.00	28.00	500.00	500.00	455.00	252.00	274.00	252.00	3179.00	1406.00	129.68	562.66	604.55	245.97	284.40	404.58	499.00
0.9600000000000001	1282.78	1344.04	1256.34	1406.07	1407.78	1305.32	120.08	83.43	82.12	211.03	153.97	145.54	411.69	412.46	380.74	674.30	670.99	646.48	39.84	40.05	37.61	10.61	11.86	11.45	1025.57	1016.38	975.80	1099.72	1146.44	1066.04	1101.07	1136.32	998.28	11.43	10.87	9.59	1707.60	1766.85	1672.44	2.20	2.28	1.09	31.43	37.11	34.33	4.78	0.00	0.00	276.09	270.15	265.31	578.33	574.86	573.49	85.30	89.25	77.93	847.56	854.66	848.82	77.15	81.35	74.18	151.66	153.16	144.74	15.67	15.88	14.48	237.74	241.25	224.12	0.46	0.13	0.25	60.59	67.46	61.77	4.98	6.51	5.31	32.00	32.00	31.00	505.00	550.00	500.00	330.00	339.00	300.00	3264.00	1508.50	185.21	705.70	704.27	282.57	356.70	458.40	555.30
0.9700000000000001	1444.23	1497.25	1441.53	1578.82	1585.02	1481.57	155.13	117.54	112.07	270.52	203.66	188.86	508.01	500.20	458.64	758.99	749.79	734.08	49.38	49.09	46.36	13.04	14.68	14.14	1163.16	1143.62	1101.16	1243.36	1308.19	1235.72	1262.71	1308.35	1162.84	16.44	15.26	13.64	1904.26	1950.83	1877.19	5.01	5.36	2.73	37.80	44.44	40.55	5.56	0.00	0.00	320.64	322.92	308.49	655.65	646.22	642.58	103.65	109.78	96.64	959.39	962.29	941.13	97.62	101.69	94.54	184.59	187.18	176.96	20.81	21.78	19.62	290.90	295.51	279.29	0.53	0.20	0.36	82.75	91.07	85.44	7.03	9.05	7.58	35.00	36.00	34.00	550.00	550.00	550.00	398.00	398.00	379.00	3424.70	1633.85	262.23	895.25	843.24	334.26	461.60	529.37	644.00
0.9800000000000001	1694.68	1772.62	1700.24	1837.93	1838.39	1739.01	221.26	166.28	165.81	363.12	282.49	266.53	668.59	660.28	596.48	885.83	868.35	853.83	62.69	63.06	60.11	17.15	19.23	18.67	1372.78	1338.79	1306.65	1458.71	1520.56	1463.19	1518.64	1558.41	1413.14	24.58	23.23	21.17	2174.34	2312.91	2165.26	12.99	13.21	7.75	48.05	56.14	51.15	6.84	0.00	0.00	414.27	407.34	392.53	775.61	765.13	748.24	132.06	143.85	125.21	1136.29	1136.25	1114.04	132.11	138.47	131.28	245.19	253.11	234.79	30.02	31.45	28.10	367.71	388.61	362.84	0.58	0.30	0.50	129.13	135.43	127.01	10.84	13.99	11.54	40.00	40.00	39.00	655.00	750.00	619.00	500.00	500.00	500.00	3632.00	1834.70	392.11	1207.69	1051.14	431.92	621.75	649.14	779.30
0.9900000000000001	2166.37	2220.37	2188.50	2326.29	2410.10	2211.64	349.35	292.54	288.49	543.71	448.13	432.74	1076.24	1059.88	956.50	1147.05	1112.66	1092.59	90.88	91.06	86.68	24.86	28.24	28.87	1806.94	1761.43	1689.07	1885.20	1919.19	1938.13	1955.61	2112.66	1905.81	44.39	43.89	38.88	2744.49	2874.65	2800.87	41.25	40.43	31.24	71.36	79.87	74.11	9.31	0.00	0.00	625.35	648.79	621.67	1026.44	1009.29	976.09	197.17	205.25	185.62	1484.99	1515.87	1459.55	215.64	231.15	215.20	393.73	408.58	372.61	53.39	56.59	49.41	577.89	616.89	563.89	0.68	0.51	0.61	239.60	240.13	249.89	20.71	25.26	21.53	48.00	48.00	46.00	1000.00	1000.00	951.00	655.00	655.00	619.00	3651.00	2216.30	654.31	1878.12	1465.10	619.69	929.64	864.34	1036.40
1.0	7376.71	8157.78	10752.56	8362.36	9667.13	14007.34	2613.31	3813.29	4169.81	3775.11	2812.04	5337.04	6431.33	7400.66	10752.56	4729.74	4557.14	4961.33	1466.03	1196.43	928.49	342.86	569.71	351.83	10643.38	7674.78	11039.91	7366.58	8133.66	8014.43	8314.76	9284.74	13950.04	628.56	544.63	516.91	8432.99	10936.73	13980.06	5900.66	5490.28	5681.54	1023.21	1265.79	1390.88	100.61	370.13	394.93	6351.44	5709.59	4003.21	4693.86	4388.73	5738.46	1678.41	1983.01	1588.53	6496.11	6466.74	5748.81	5459.56	5800.93	4309.29	4630.23	3470.38	5645.86	1351.11	1136.08	1394.89	5459.63	6745.76	5957.14	19.76	21.33	6.23	3965.69	4747.91	4100.38	1344.14	1495.94	1209.86	307.00	138.00	196.00	4010.00	4010.00	4449.00	4010.00	4010.00	4449.00	4321.00	37762.50	8062.30	15646.39	12768.70	4862.62	8254.62	12808.62	14344.50

	mobile_number	circle_id	last_date_of_month_6	last_date_of_month_7	last_date_of_month_8	last_date_of_month_9	arpu_6	arpu_7	arpu_8	arpu_9	onnet_mou_6	onnet_mou_7	onnet_mou_8	onnet_mou_9	offnet_mou_6	offnet_mou_7	offnet_mou_8	offnet_mou_9	roam_ic_mou_6	roam_ic_mou_7	roam_ic_mou_8	roam_ic_mou_9	roam_og_mou_6	roam_og_mou_7	roam_og_mou_8	roam_og_mou_9	loc_og_t2t_mou_6	loc_og_t2t_mou_7	loc_og_t2t_mou_8	loc_og_t2t_mou_9	loc_og_t2m_mou_6	loc_og_t2m_mou_7	loc_og_t2m_mou_8	loc_og_t2m_mou_9	loc_og_t2f_mou_6	loc_og_t2f_mou_7	loc_og_t2f_mou_8	loc_og_t2f_mou_9	loc_og_t2c_mou_6	loc_og_t2c_mou_7	loc_og_t2c_mou_8	loc_og_t2c_mou_9	loc_og_mou_6	loc_og_mou_7	loc_og_mou_8	loc_og_mou_9	std_og_t2t_mou_6	std_og_t2t_mou_7	std_og_t2t_mou_8	std_og_t2t_mou_9	std_og_t2m_mou_6	std_og_t2m_mou_7	std_og_t2m_mou_8	std_og_t2m_mou_9	std_og_t2f_mou_6	std_og_t2f_mou_7	std_og_t2f_mou_8	std_og_t2f_mou_9	std_og_t2c_mou_6	std_og_t2c_mou_7	std_og_t2c_mou_9	std_og_mou_6	std_og_mou_7	std_og_mou_8	std_og_mou_9	isd_og_mou_6	isd_og_mou_7	isd_og_mou_9	spl_og_mou_6	spl_og_mou_7	spl_og_mou_8	spl_og_mou_9	og_others_6	og_others_7	og_others_9	total_og_mou_6	total_og_mou_7	total_og_mou_8	total_og_mou_9	loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	loc_ic_t2t_mou_8	loc_ic_t2t_mou_9	loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	loc_ic_t2m_mou_9	loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	loc_ic_t2f_mou_9	loc_ic_mou_6	loc_ic_mou_7	loc_ic_mou_8	loc_ic_mou_9	std_ic_t2t_mou_6	std_ic_t2t_mou_7	std_ic_t2t_mou_8	std_ic_t2t_mou_9	std_ic_t2m_mou_6	std_ic_t2m_mou_7	std_ic_t2m_mou_8	std_ic_t2m_mou_9	std_ic_t2f_mou_6	std_ic_t2f_mou_7	std_ic_t2f_mou_8	std_ic_t2f_mou_9	std_ic_t2o_mou_6	std_ic_t2o_mou_7	std_ic_t2o_mou_9	std_ic_mou_6	std_ic_mou_7	std_ic_mou_8	std_ic_mou_9	total_ic_mou_6	total_ic_mou_7	total_ic_mou_8	total_ic_mou_9	spl_ic_mou_6	spl_ic_mou_7	spl_ic_mou_9	isd_ic_mou_6	isd_ic_mou_7	isd_ic_mou_9	ic_others_6	ic_others_7	ic_others_9	total_rech_num_6	total_rech_num_7	total_rech_num_8	total_rech_num_9	total_rech_amt_6	total_rech_amt_7	total_rech_amt_8	total_rech_amt_9	max_rech_amt_6	max_rech_amt_7	max_rech_amt_8	max_rech_amt_9	date_of_last_rech_6	date_of_last_rech_7	date_of_last_rech_8	date_of_last_rech_9	last_day_rch_amt_6	last_day_rch_amt_7	last_day_rch_amt_8	last_day_rch_amt_9	date_of_last_rech_data_6	date_of_last_rech_data_7	date_of_last_rech_data_8	date_of_last_rech_data_9	total_rech_data_6	total_rech_data_7	total_rech_data_8	total_rech_data_9	max_rech_data_6	max_rech_data_7	max_rech_data_8	max_rech_data_9	count_rech_2g_6	count_rech_2g_7	count_rech_2g_8	count_rech_2g_9	count_rech_3g_6	count_rech_3g_7	count_rech_3g_8	count_rech_3g_9	av_rech_amt_data_6	av_rech_amt_data_7	av_rech_amt_data_8	av_rech_amt_data_9	vol_2g_mb_6	vol_2g_mb_7	vol_2g_mb_8	vol_3g_mb_6	vol_3g_mb_7	vol_3g_mb_8	vol_3g_mb_9	arpu_3g_6	arpu_3g_7	arpu_3g_8	arpu_3g_9	arpu_2g_6	arpu_2g_7	arpu_2g_8	arpu_2g_9	night_pck_user_6	night_pck_user_7	night_pck_user_8	night_pck_user_9	monthly_2g_7	sachet_2g_6	sachet_2g_8	sachet_2g_9	monthly_3g_6	monthly_3g_7	monthly_3g_8	fb_user_6	fb_user_7	fb_user_8	fb_user_9	aon	aug_vbc_3g	jun_vbc_3g	sep_vbc_3g
0	7000842753	109	6/30/2014	7/31/2014	8/31/2014	9/30/2014	197.385	214.816	213.803	21.100	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	NaN	NaN	NaN	0.00	NaN	NaN	NaN	NaN	NaN	NaN	0.00	NaN	NaN	NaN	NaN	0.00	0.00	0.00	0.00	NaN	NaN	0.16	NaN	NaN	NaN	4.13	NaN	NaN	NaN	1.15	NaN	NaN	NaN	5.44	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	NaN	NaN	NaN	0.00	NaN	0.00	0.00	5.44	0.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	4	3	2	6	362	252	252	0	252	252	252	0	6/21/2014	7/16/2014	8/8/2014	9/28/2014	252	252	252	0	6/21/2014	7/16/2014	8/8/2014	NaN	1.0	1.0	1.0	NaN	252.0	252.0	252.0	NaN	0.0	0.0	0.0	NaN	1.0	1.0	1.0	NaN	252.0	252.0	252.0	NaN	30.13	1.32	5.75	83.57	150.76	109.61	0.00	212.17	212.17	212.17	NaN	212.17	212.17	212.17	NaN	0.0	0.0	0.0	NaN	0	0	0	0	1	1	1	1.0	1.0	1.0	NaN	968	30.4	101.20	3.58
1	7001865778	109	6/30/2014	7/31/2014	8/31/2014	9/30/2014	34.047	355.074	268.321	86.285	24.11	78.68	7.68	18.34	15.74	99.84	304.76	53.76	0.0	0.00	0.00	0.00	0.0	0.00	0.00	0.00	23.88	74.56	7.68	18.34	11.51	75.94	291.86	53.76	0.00	0.00	0.00	0.00	0.0	2.91	0.00	0.00	35.39	150.51	299.54	72.11	0.23	4.11	0.00	0.00	0.00	0.46	0.13	0.00	0.00	0.00	0.00	0.0	0.0	0.0	0.0	0.23	4.58	0.13	0.00	0.0	0.0	0.0	4.68	23.43	12.76	0.00	0.00	0.0	0.0	40.31	178.53	312.44	72.11	1.61	29.91	29.23	116.09	17.48	65.38	375.58	56.93	0.00	8.93	3.61	0.00	19.09	104.23	408.43	173.03	0.00	0.00	2.35	0.00	5.90	0.00	12.49	15.01	0.00	0.00	0.00	0.00	0.0	0.0	0.0	5.90	0.00	14.84	15.01	26.83	104.23	423.28	188.04	0.00	0.0	0.00	1.83	0.00	0.00	0.00	0.00	0.00	4	9	11	5	74	384	283	121	44	154	65	50	6/29/2014	7/31/2014	8/28/2014	9/30/2014	44	23	30	0	NaN	7/25/2014	8/10/2014	NaN	NaN	1.0	2.0	NaN	NaN	154.0	25.0	NaN	NaN	1.0	2.0	NaN	NaN	0.0	0.0	NaN	NaN	154.0	50.0	NaN	0.00	108.07	365.47	0.00	0.00	0.00	0.00	NaN	0.00	0.00	NaN	NaN	28.61	7.60	NaN	NaN	0.0	0.0	NaN	1	0	2	0	0	0	0	NaN	1.0	1.0	NaN	1006	0.0	0.00	0.00
2	7001625959	109	6/30/2014	7/31/2014	8/31/2014	9/30/2014	167.690	189.058	210.226	290.714	11.54	55.24	37.26	74.81	143.33	220.59	208.36	118.91	0.0	0.00	0.00	38.49	0.0	0.00	0.00	70.94	7.19	28.74	13.58	14.39	29.34	16.86	38.46	28.16	24.11	21.79	15.61	22.24	0.0	135.54	45.76	0.48	60.66	67.41	67.66	64.81	4.34	26.49	22.58	8.76	41.81	67.41	75.53	9.28	1.48	14.76	22.83	0.0	0.0	0.0	0.0	47.64	108.68	120.94	18.04	0.0	0.0	0.0	46.56	236.84	96.84	42.08	0.45	0.0	0.0	155.33	412.94	285.46	124.94	115.69	71.11	67.46	148.23	14.38	15.44	38.89	38.98	99.48	122.29	49.63	158.19	229.56	208.86	155.99	345.41	72.41	71.29	28.69	49.44	45.18	177.01	167.09	118.18	21.73	58.34	43.23	3.86	0.0	0.0	0.0	139.33	306.66	239.03	171.49	370.04	519.53	395.03	517.74	0.21	0.0	0.45	0.00	0.85	0.01	0.93	3.14	0.36	5	4	2	7	168	315	116	358	86	200	86	100	6/17/2014	7/24/2014	8/14/2014	9/29/2014	0	200	86	0	NaN	NaN	NaN	9/17/2014	NaN	NaN	NaN	1.0	NaN	NaN	NaN	46.0	NaN	NaN	NaN	1.0	NaN	NaN	NaN	0.0	NaN	NaN	NaN	46.0	0.00	0.00	0.00	0.00	0.00	0.00	8.42	NaN	NaN	NaN	2.84	NaN	NaN	NaN	0.0	NaN	NaN	NaN	0.0	0	0	0	1	0	0	0	NaN	NaN	NaN	1.0	1103	0.0	4.17	0.00
3	7001204172	109	6/30/2014	7/31/2014	8/31/2014	9/30/2014	221.338	251.102	508.054	389.500	99.91	54.39	310.98	241.71	123.31	109.01	71.68	113.54	0.0	54.86	44.38	0.00	0.0	28.09	39.04	0.00	73.68	34.81	10.61	15.49	107.43	83.21	22.46	65.46	1.91	0.65	4.91	2.06	0.0	0.00	0.00	0.00	183.03	118.68	37.99	83.03	26.23	14.89	289.58	226.21	2.99	1.73	6.53	9.99	0.00	0.00	0.00	0.0	0.0	0.0	0.0	29.23	16.63	296.11	236.21	0.0	0.0	0.0	10.96	0.00	18.09	43.29	0.00	0.0	0.0	223.23	135.31	352.21	362.54	62.08	19.98	8.04	41.73	113.96	64.51	20.28	52.86	57.43	27.09	19.84	65.59	233.48	111.59	48.18	160.19	43.48	66.44	0.00	129.84	1.33	38.56	4.94	13.98	1.18	0.00	0.00	0.00	0.0	0.0	0.0	45.99	105.01	4.94	143.83	280.08	216.61	53.13	305.38	0.59	0.0	0.55	0.00	0.00	0.00	0.00	0.00	0.80	10	11	18	14	230	310	601	410	60	50	50	50	6/28/2014	7/31/2014	8/31/2014	9/30/2014	30	50	50	30	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.00	0.00	0.00	0.00	0.00	0.00	0.00	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	0	0	0	0	0	0	NaN	NaN	NaN	NaN	2491	0.0	0.00	0.00
4	7000142493	109	6/30/2014	7/31/2014	8/31/2014	9/30/2014	261.636	309.876	238.174	163.426	50.31	149.44	83.89	58.78	76.96	91.88	124.26	45.81	0.0	0.00	0.00	0.00	0.0	0.00	0.00	0.00	50.31	149.44	83.89	58.78	67.64	91.88	124.26	37.89	0.00	0.00	0.00	1.93	0.0	0.00	0.00	0.00	117.96	241.33	208.16	98.61	0.00	0.00	0.00	0.00	9.31	0.00	0.00	0.00	0.00	0.00	0.00	0.0	0.0	0.0	0.0	9.31	0.00	0.00	0.00	0.0	0.0	0.0	0.00	0.00	0.00	5.98	0.00	0.0	0.0	127.28	241.33	208.16	104.59	105.68	88.49	233.81	154.56	106.84	109.54	104.13	48.24	1.50	0.00	0.00	0.00	214.03	198.04	337.94	202.81	0.00	0.00	0.86	2.31	1.93	0.25	0.00	0.00	0.00	0.00	0.00	0.00	0.0	0.0	0.0	1.93	0.25	0.86	2.31	216.44	198.29	338.81	205.31	0.00	0.0	0.18	0.00	0.00	0.00	0.48	0.00	0.00	5	6	3	4	196	350	287	200	56	110	110	50	6/26/2014	7/28/2014	8/9/2014	9/28/2014	50	110	110	50	6/4/2014	NaN	NaN	NaN	1.0	NaN	NaN	NaN	56.0	NaN	NaN	NaN	1.0	NaN	NaN	NaN	0.0	NaN	NaN	NaN	56.0	NaN	NaN	NaN	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	NaN	NaN	NaN	0.00	NaN	NaN	NaN	0.0	NaN	NaN	NaN	0	1	0	0	0	0	0	0.0	NaN	NaN	NaN	1526	0.0	0.00	0.00

	Datatype	Non_Null_Count	Null_Count	Null_Percentage	Unique_Values_Count
arpu_3g_6	float64	25153	74846	74.85	7418
night_pck_user_6	float64	25153	74846	74.85	2
total_rech_data_6	float64	25153	74846	74.85	37
arpu_2g_6	float64	25153	74846	74.85	6990
max_rech_data_6	float64	25153	74846	74.85	48
fb_user_6	float64	25153	74846	74.85	2
av_rech_amt_data_6	float64	25153	74846	74.85	887
date_of_last_rech_data_6	object	25153	74846	74.85	30
count_rech_2g_6	float64	25153	74846	74.85	31
count_rech_3g_6	float64	25153	74846	74.85	25
date_of_last_rech_data_7	object	25571	74428	74.43	31
total_rech_data_7	float64	25571	74428	74.43	42
fb_user_7	float64	25571	74428	74.43	2
max_rech_data_7	float64	25571	74428	74.43	48
night_pck_user_7	float64	25571	74428	74.43	2
count_rech_2g_7	float64	25571	74428	74.43	36
av_rech_amt_data_7	float64	25571	74428	74.43	961
arpu_2g_7	float64	25571	74428	74.43	6586
count_rech_3g_7	float64	25571	74428	74.43	28
arpu_3g_7	float64	25571	74428	74.43	7246
total_rech_data_9	float64	25922	74077	74.08	37
count_rech_3g_9	float64	25922	74077	74.08	27
fb_user_9	float64	25922	74077	74.08	2
max_rech_data_9	float64	25922	74077	74.08	50
arpu_3g_9	float64	25922	74077	74.08	8063
date_of_last_rech_data_9	object	25922	74077	74.08	30
night_pck_user_9	float64	25922	74077	74.08	2
arpu_2g_9	float64	25922	74077	74.08	6795
count_rech_2g_9	float64	25922	74077	74.08	32
av_rech_amt_data_9	float64	25922	74077	74.08	945
total_rech_data_8	float64	26339	73660	73.66	46
arpu_3g_8	float64	26339	73660	73.66	7787
fb_user_8	float64	26339	73660	73.66	2
night_pck_user_8	float64	26339	73660	73.66	2
av_rech_amt_data_8	float64	26339	73660	73.66	973
max_rech_data_8	float64	26339	73660	73.66	50
count_rech_3g_8	float64	26339	73660	73.66	29
arpu_2g_8	float64	26339	73660	73.66	6652
count_rech_2g_8	float64	26339	73660	73.66	34
date_of_last_rech_data_8	object	26339	73660	73.66	31
ic_others_9	float64	92254	7745	7.75	1923
std_og_mou_9	float64	92254	7745	7.75	26553
std_og_t2c_mou_9	float64	92254	7745	7.75	1
isd_ic_mou_9	float64	92254	7745	7.75	5557
std_ic_mou_9	float64	92254	7745	7.75	11266
isd_og_mou_9	float64	92254	7745	7.75	1255
spl_og_mou_9	float64	92254	7745	7.75	4095
spl_ic_mou_9	float64	92254	7745	7.75	384
og_others_9	float64	92254	7745	7.75	235
loc_ic_t2t_mou_9	float64	92254	7745	7.75	12993
std_ic_t2o_mou_9	float64	92254	7745	7.75	1
loc_ic_t2m_mou_9	float64	92254	7745	7.75	21484
std_ic_t2f_mou_9	float64	92254	7745	7.75	3090
loc_ic_t2f_mou_9	float64	92254	7745	7.75	7091
loc_ic_mou_9	float64	92254	7745	7.75	27697
std_ic_t2m_mou_9	float64	92254	7745	7.75	8933
std_og_t2f_mou_9	float64	92254	7745	7.75	2295
std_og_t2t_mou_9	float64	92254	7745	7.75	17934
std_ic_t2t_mou_9	float64	92254	7745	7.75	6157
loc_og_mou_9	float64	92254	7745	7.75	25376
roam_og_mou_9	float64	92254	7745	7.75	5882
loc_og_t2m_mou_9	float64	92254	7745	7.75	20141
loc_og_t2f_mou_9	float64	92254	7745	7.75	3758
roam_ic_mou_9	float64	92254	7745	7.75	4827
offnet_mou_9	float64	92254	7745	7.75	30077
loc_og_t2c_mou_9	float64	92254	7745	7.75	2332
loc_og_t2t_mou_9	float64	92254	7745	7.75	12949
std_og_t2m_mou_9	float64	92254	7745	7.75	19052
onnet_mou_9	float64	92254	7745	7.75	23565
onnet_mou_8	float64	94621	5378	5.38	24089
std_ic_t2t_mou_8	float64	94621	5378	5.38	6352
std_ic_mou_8	float64	94621	5378	5.38	11662
loc_ic_t2t_mou_8	float64	94621	5378	5.38	13346
roam_og_mou_8	float64	94621	5378	5.38	6504
std_ic_t2m_mou_8	float64	94621	5378	5.38	9304
loc_ic_mou_8	float64	94621	5378	5.38	28200
std_ic_t2f_mou_8	float64	94621	5378	5.38	3051
roam_ic_mou_8	float64	94621	5378	5.38	5315
std_ic_t2o_mou_8	float64	94621	5378	5.38	1
loc_og_t2t_mou_8	float64	94621	5378	5.38	13336
loc_ic_t2f_mou_8	float64	94621	5378	5.38	7097
offnet_mou_8	float64	94621	5378	5.38	30908
loc_ic_t2m_mou_8	float64	94621	5378	5.38	21886
loc_og_t2m_mou_8	float64	94621	5378	5.38	20544
isd_og_mou_8	float64	94621	5378	5.38	1276
ic_others_8	float64	94621	5378	5.38	1896
og_others_8	float64	94621	5378	5.38	216
spl_ic_mou_8	float64	94621	5378	5.38	102
loc_og_t2f_mou_8	float64	94621	5378	5.38	3807
std_og_t2m_mou_8	float64	94621	5378	5.38	19786
spl_og_mou_8	float64	94621	5378	5.38	4390
std_og_t2c_mou_8	float64	94621	5378	5.38	1
isd_ic_mou_8	float64	94621	5378	5.38	5844
loc_og_t2c_mou_8	float64	94621	5378	5.38	2516
std_og_t2f_mou_8	float64	94621	5378	5.38	2333
std_og_t2t_mou_8	float64	94621	5378	5.38	18291
loc_og_mou_8	float64	94621	5378	5.38	25990
std_og_mou_8	float64	94621	5378	5.38	27491
date_of_last_rech_9	object	95239	4760	4.76	30
std_ic_t2f_mou_6	float64	96062	3937	3.94	3125
ic_others_6	float64	96062	3937	3.94	1817
isd_ic_mou_6	float64	96062	3937	3.94	5521
std_ic_t2m_mou_6	float64	96062	3937	3.94	9308
std_ic_mou_6	float64	96062	3937	3.94	11646
spl_ic_mou_6	float64	96062	3937	3.94	84
std_ic_t2o_mou_6	float64	96062	3937	3.94	1
loc_ic_t2f_mou_6	float64	96062	3937	3.94	7250
loc_ic_t2t_mou_6	float64	96062	3937	3.94	13540
std_og_t2c_mou_6	float64	96062	3937	3.94	1
std_og_t2f_mou_6	float64	96062	3937	3.94	2450
std_og_mou_6	float64	96062	3937	3.94	27502
std_og_t2m_mou_6	float64	96062	3937	3.94	19734
isd_og_mou_6	float64	96062	3937	3.94	1381
std_og_t2t_mou_6	float64	96062	3937	3.94	18244
spl_og_mou_6	float64	96062	3937	3.94	3965
loc_og_mou_6	float64	96062	3937	3.94	26372
og_others_6	float64	96062	3937	3.94	1018
loc_og_t2c_mou_6	float64	96062	3937	3.94	2235
loc_og_t2m_mou_6	float64	96062	3937	3.94	20905
loc_og_t2f_mou_6	float64	96062	3937	3.94	3860
loc_og_t2t_mou_6	float64	96062	3937	3.94	13539
roam_og_mou_6	float64	96062	3937	3.94	8038
std_ic_t2t_mou_6	float64	96062	3937	3.94	6279
onnet_mou_6	float64	96062	3937	3.94	24313
loc_ic_mou_6	float64	96062	3937	3.94	28569
offnet_mou_6	float64	96062	3937	3.94	31140
roam_ic_mou_6	float64	96062	3937	3.94	6512
loc_ic_t2m_mou_6	float64	96062	3937	3.94	22065
loc_og_t2c_mou_7	float64	96140	3859	3.86	2426
roam_ic_mou_7	float64	96140	3859	3.86	5230
loc_og_mou_7	float64	96140	3859	3.86	26091
loc_og_t2t_mou_7	float64	96140	3859	3.86	13411
offnet_mou_7	float64	96140	3859	3.86	31023
loc_og_t2f_mou_7	float64	96140	3859	3.86	3863
std_og_t2t_mou_7	float64	96140	3859	3.86	18567
std_ic_t2t_mou_7	float64	96140	3859	3.86	6481
onnet_mou_7	float64	96140	3859	3.86	24336
std_og_t2m_mou_7	float64	96140	3859	3.86	20018
loc_og_t2m_mou_7	float64	96140	3859	3.86	20637
std_og_t2f_mou_7	float64	96140	3859	3.86	2391
roam_og_mou_7	float64	96140	3859	3.86	6639
std_og_t2c_mou_7	float64	96140	3859	3.86	1
std_ic_t2m_mou_7	float64	96140	3859	3.86	9464
isd_og_mou_7	float64	96140	3859	3.86	1380
ic_others_7	float64	96140	3859	3.86	2002
loc_ic_t2f_mou_7	float64	96140	3859	3.86	7395
loc_ic_t2m_mou_7	float64	96140	3859	3.86	21918
std_ic_mou_7	float64	96140	3859	3.86	11889
loc_ic_t2t_mou_7	float64	96140	3859	3.86	13511
std_ic_t2f_mou_7	float64	96140	3859	3.86	3209
loc_ic_mou_7	float64	96140	3859	3.86	28390
spl_ic_mou_7	float64	96140	3859	3.86	107
og_others_7	float64	96140	3859	3.86	187
spl_og_mou_7	float64	96140	3859	3.86	4396
isd_ic_mou_7	float64	96140	3859	3.86	5789
std_ic_t2o_mou_7	float64	96140	3859	3.86	1
std_og_mou_7	float64	96140	3859	3.86	27951
date_of_last_rech_8	object	96377	3622	3.62	31
date_of_last_rech_7	object	98232	1767	1.77	31
last_date_of_month_9	object	98340	1659	1.66	1
date_of_last_rech_6	object	98392	1607	1.61	30
last_date_of_month_8	object	98899	1100	1.10	1
loc_ic_t2o_mou	float64	98981	1018	1.02	1
std_og_t2o_mou	float64	98981	1018	1.02	1
loc_og_t2o_mou	float64	98981	1018	1.02	1
last_date_of_month_7	object	99398	601	0.60	1
sachet_3g_8	int64	99999	0	0.00	29
jul_vbc_3g	float64	99999	0	0.00	14162
aug_vbc_3g	float64	99999	0	0.00	14676
aon	int64	99999	0	0.00	3489
jun_vbc_3g	float64	99999	0	0.00	13312
monthly_2g_9	int64	99999	0	0.00	5
sachet_3g_6	int64	99999	0	0.00	25
vol_3g_mb_9	float64	99999	0	0.00	14472
sachet_3g_7	int64	99999	0	0.00	27
monthly_2g_8	int64	99999	0	0.00	6
monthly_3g_9	int64	99999	0	0.00	11
monthly_3g_8	int64	99999	0	0.00	12
sachet_3g_9	int64	99999	0	0.00	27
monthly_3g_7	int64	99999	0	0.00	15
monthly_3g_6	int64	99999	0	0.00	12
sachet_2g_9	int64	99999	0	0.00	32
sachet_2g_8	int64	99999	0	0.00	34
sachet_2g_7	int64	99999	0	0.00	35
sachet_2g_6	int64	99999	0	0.00	32
monthly_2g_7	int64	99999	0	0.00	6
monthly_2g_6	int64	99999	0	0.00	5
mobile_number	int64	99999	0	0.00	99999
vol_3g_mb_8	float64	99999	0	0.00	14960
total_og_mou_9	float64	99999	0	0.00	39160
total_rech_num_7	int64	99999	0	0.00	101
total_rech_num_6	int64	99999	0	0.00	102
total_ic_mou_9	float64	99999	0	0.00	31260
total_ic_mou_8	float64	99999	0	0.00	32128
total_ic_mou_7	float64	99999	0	0.00	32242
total_ic_mou_6	float64	99999	0	0.00	32247
circle_id	int64	99999	0	0.00	1
total_og_mou_8	float64	99999	0	0.00	40074
vol_3g_mb_7	float64	99999	0	0.00	14519
total_og_mou_7	float64	99999	0	0.00	40477
total_og_mou_6	float64	99999	0	0.00	40327
arpu_9	float64	99999	0	0.00	79937
arpu_8	float64	99999	0	0.00	83615
arpu_7	float64	99999	0	0.00	85308
arpu_6	float64	99999	0	0.00	85681
last_date_of_month_6	object	99999	0	0.00	1
total_rech_num_8	int64	99999	0	0.00	96
total_rech_num_9	int64	99999	0	0.00	97
total_rech_amt_6	int64	99999	0	0.00	2305
total_rech_amt_7	int64	99999	0	0.00	2329
vol_3g_mb_6	float64	99999	0	0.00	13773
vol_2g_mb_9	float64	99999	0	0.00	13919
vol_2g_mb_8	float64	99999	0	0.00	14994
vol_2g_mb_7	float64	99999	0	0.00	15114
vol_2g_mb_6	float64	99999	0	0.00	15201
last_day_rch_amt_9	int64	99999	0	0.00	185
last_day_rch_amt_8	int64	99999	0	0.00	199
last_day_rch_amt_7	int64	99999	0	0.00	173
last_day_rch_amt_6	int64	99999	0	0.00	186
max_rech_amt_9	int64	99999	0	0.00	201
max_rech_amt_8	int64	99999	0	0.00	213
max_rech_amt_7	int64	99999	0	0.00	183
max_rech_amt_6	int64	99999	0	0.00	202
total_rech_amt_9	int64	99999	0	0.00	2304
total_rech_amt_8	int64	99999	0	0.00	2347
sep_vbc_3g	float64	99999	0	0.00	3720

	Datatype	Non_Null_Count	Null_Count	Null_Percentage	Unique_Values_Count
std_ic_t2o_mou_9	category	28307	1704	5.68	1
spl_og_mou_9	float64	28307	1704	5.68	2966
isd_og_mou_9	float64	28307	1704	5.68	908
roam_ic_mou_9	float64	28307	1704	5.68	3370
std_og_mou_9	float64	28307	1704	5.68	15900
roam_og_mou_9	float64	28307	1704	5.68	4004
std_ic_t2f_mou_9	float64	28307	1704	5.68	1971
std_og_t2c_mou_9	category	28307	1704	5.68	1
loc_og_t2t_mou_9	float64	28307	1704	5.68	10360
std_og_t2f_mou_9	float64	28307	1704	5.68	1595
std_ic_mou_9	float64	28307	1704	5.68	7745
loc_og_t2m_mou_9	float64	28307	1704	5.68	15585
std_og_t2m_mou_9	float64	28307	1704	5.68	12445
loc_og_t2f_mou_9	float64	28307	1704	5.68	3111
std_og_t2t_mou_9	float64	28307	1704	5.68	11141
loc_ic_mou_9	float64	28307	1704	5.68	18018
loc_og_t2c_mou_9	float64	28307	1704	5.68	1576
offnet_mou_9	float64	28307	1704	5.68	20452
loc_og_mou_9	float64	28307	1704	5.68	18207
spl_ic_mou_9	float64	28307	1704	5.68	287
std_ic_t2m_mou_9	float64	28307	1704	5.68	6168
loc_ic_t2f_mou_9	float64	28307	1704	5.68	4611
ic_others_9	float64	28307	1704	5.68	1284
loc_ic_t2m_mou_9	float64	28307	1704	5.68	15194
loc_ic_t2t_mou_9	float64	28307	1704	5.68	9407
std_ic_t2t_mou_9	float64	28307	1704	5.68	4280
isd_ic_mou_9	float64	28307	1704	5.68	3329
og_others_9	float64	28307	1704	5.68	132
onnet_mou_9	float64	28307	1704	5.68	16674
std_og_mou_8	float64	29073	938	3.13	16864
std_og_t2m_mou_8	float64	29073	938	3.13	13326
og_others_8	float64	29073	938	3.13	133
loc_ic_t2f_mou_8	float64	29073	938	3.13	4705
std_og_t2t_mou_8	float64	29073	938	3.13	11781
loc_og_mou_8	float64	29073	938	3.13	18885
std_ic_t2o_mou_8	category	29073	938	3.13	1
loc_ic_t2m_mou_8	float64	29073	938	3.13	15598
std_ic_t2m_mou_8	float64	29073	938	3.13	6420
std_ic_t2t_mou_8	float64	29073	938	3.13	4486
std_og_t2f_mou_8	float64	29073	938	3.13	1627
std_ic_t2f_mou_8	float64	29073	938	3.13	1941
spl_og_mou_8	float64	29073	938	3.13	3238
loc_ic_t2t_mou_8	float64	29073	938	3.13	9671
std_og_t2c_mou_8	category	29073	938	3.13	1
isd_og_mou_8	float64	29073	938	3.13	940
loc_ic_mou_8	float64	29073	938	3.13	18573
roam_ic_mou_8	float64	29073	938	3.13	3655
isd_ic_mou_8	float64	29073	938	3.13	3493
onnet_mou_8	float64	29073	938	3.13	17604
loc_og_t2c_mou_8	float64	29073	938	3.13	1730
spl_ic_mou_8	float64	29073	938	3.13	85
loc_og_t2f_mou_8	float64	29073	938	3.13	3124
std_ic_mou_8	float64	29073	938	3.13	8033
roam_og_mou_8	float64	29073	938	3.13	4382
ic_others_8	float64	29073	938	3.13	1259
loc_og_t2m_mou_8	float64	29073	938	3.13	16165
loc_og_t2t_mou_8	float64	29073	938	3.13	10772
offnet_mou_8	float64	29073	938	3.13	21513
date_of_last_rech_9	datetime64[ns]	29145	866	2.89	30
date_of_last_rech_8	datetime64[ns]	29417	594	1.98	31
last_date_of_month_9	object	29651	360	1.20	1
std_ic_mou_6	float64	29695	316	1.05	8391
offnet_mou_6	float64	29695	316	1.05	22454
std_ic_t2f_mou_6	float64	29695	316	1.05	2033
isd_ic_mou_6	float64	29695	316	1.05	3429
ic_others_6	float64	29695	316	1.05	1227
onnet_mou_6	float64	29695	316	1.05	18813
std_ic_t2m_mou_6	float64	29695	316	1.05	6680
loc_ic_t2t_mou_6	float64	29695	316	1.05	9872
loc_ic_t2m_mou_6	float64	29695	316	1.05	16015
loc_ic_t2f_mou_6	float64	29695	316	1.05	4817
loc_ic_mou_6	float64	29695	316	1.05	19133
std_ic_t2t_mou_6	float64	29695	316	1.05	4608
og_others_6	float64	29695	316	1.05	862
spl_og_mou_6	float64	29695	316	1.05	3053
roam_ic_mou_6	float64	29695	316	1.05	4338
spl_ic_mou_6	float64	29695	316	1.05	78
std_og_t2t_mou_6	float64	29695	316	1.05	12777
loc_og_t2c_mou_6	float64	29695	316	1.05	1658
std_og_t2m_mou_6	float64	29695	316	1.05	14518
loc_og_t2f_mou_6	float64	29695	316	1.05	3252
std_og_t2f_mou_6	float64	29695	316	1.05	1773
loc_og_t2m_mou_6	float64	29695	316	1.05	16747
std_ic_t2o_mou_6	category	29695	316	1.05	1
std_og_t2c_mou_6	category	29695	316	1.05	1
std_og_mou_6	float64	29695	316	1.05	18325
loc_og_t2t_mou_6	float64	29695	316	1.05	11151
isd_og_mou_6	float64	29695	316	1.05	1113
roam_og_mou_6	float64	29695	316	1.05	5174
loc_og_mou_6	float64	29695	316	1.05	19691
isd_ic_mou_7	float64	29708	303	1.01	3639
std_ic_t2f_mou_7	float64	29708	303	1.01	2075
std_ic_t2m_mou_7	float64	29708	303	1.01	6747
std_ic_t2o_mou_7	category	29708	303	1.01	1
ic_others_7	float64	29708	303	1.01	1371
spl_ic_mou_7	float64	29708	303	1.01	93
std_ic_t2t_mou_7	float64	29708	303	1.01	4706
std_ic_mou_7	float64	29708	303	1.01	8543
loc_ic_t2f_mou_7	float64	29708	303	1.01	4897
og_others_7	float64	29708	303	1.01	123
loc_ic_mou_7	float64	29708	303	1.01	19030
std_og_t2f_mou_7	float64	29708	303	1.01	1714
onnet_mou_7	float64	29708	303	1.01	18938
roam_ic_mou_7	float64	29708	303	1.01	3649
roam_og_mou_7	float64	29708	303	1.01	4431
loc_og_t2t_mou_7	float64	29708	303	1.01	11154
loc_og_t2m_mou_7	float64	29708	303	1.01	16872
loc_og_t2f_mou_7	float64	29708	303	1.01	3267
loc_og_t2c_mou_7	float64	29708	303	1.01	1750
loc_og_mou_7	float64	29708	303	1.01	19880
std_og_t2t_mou_7	float64	29708	303	1.01	12983
std_og_t2m_mou_7	float64	29708	303	1.01	14589
offnet_mou_7	float64	29708	303	1.01	22650
std_og_t2c_mou_7	category	29708	303	1.01	1
loc_ic_t2t_mou_7	float64	29708	303	1.01	9961
isd_og_mou_7	float64	29708	303	1.01	1125
spl_og_mou_7	float64	29708	303	1.01	3399
std_og_mou_7	float64	29708	303	1.01	18445
loc_ic_t2m_mou_7	float64	29708	303	1.01	16068
last_date_of_month_8	object	29854	157	0.52	1
std_og_t2o_mou	category	29897	114	0.38	1
loc_ic_t2o_mou	category	29897	114	0.38	1
date_of_last_rech_7	datetime64[ns]	29897	114	0.38	31
loc_og_t2o_mou	category	29897	114	0.38	1
date_of_last_rech_6	datetime64[ns]	29949	62	0.21	30
last_date_of_month_7	object	29980	31	0.10	1
sachet_3g_6	category	30011	0	0.00	25
monthly_2g_8	category	30011	0	0.00	6
vol_2g_mb_8	float64	30011	0	0.00	7310
vol_2g_mb_9	float64	30011	0	0.00	6984
vol_2g_mb_6	float64	30011	0	0.00	7809
sachet_3g_9	category	30011	0	0.00	27
sachet_3g_8	category	30011	0	0.00	29
monthly_3g_9	category	30011	0	0.00	11
vol_3g_mb_6	float64	30011	0	0.00	7043
vol_3g_mb_7	float64	30011	0	0.00	7440
vol_3g_mb_8	float64	30011	0	0.00	7151
vol_3g_mb_9	float64	30011	0	0.00	7016
monthly_2g_6	category	30011	0	0.00	5
monthly_2g_7	category	30011	0	0.00	6
monthly_2g_9	category	30011	0	0.00	5
sachet_3g_7	category	30011	0	0.00	27
sachet_2g_6	category	30011	0	0.00	30
sachet_2g_7	category	30011	0	0.00	34
sachet_2g_8	category	30011	0	0.00	34
sachet_2g_9	category	30011	0	0.00	29
vbc_3g_9	float64	30011	0	0.00	2171
monthly_3g_8	category	30011	0	0.00	12
monthly_3g_7	category	30011	0	0.00	15
vbc_3g_6	float64	30011	0	0.00	6864
vbc_3g_7	float64	30011	0	0.00	7318
vbc_3g_8	float64	30011	0	0.00	7291
aon	int64	30011	0	0.00	3321
monthly_3g_6	category	30011	0	0.00	12
vol_2g_mb_7	float64	30011	0	0.00	7813
circle_id	category	30011	0	0.00	1
last_day_rch_amt_9	int64	30011	0	0.00	170
last_day_rch_amt_8	int64	30011	0	0.00	179
last_date_of_month_6	object	30011	0	0.00	1
arpu_6	float64	30011	0	0.00	29261
arpu_7	float64	30011	0	0.00	29260
arpu_8	float64	30011	0	0.00	28405
arpu_9	float64	30011	0	0.00	27327
total_og_mou_6	float64	30011	0	0.00	24607
total_og_mou_7	float64	30011	0	0.00	24913
total_og_mou_8	float64	30011	0	0.00	23644
total_og_mou_9	float64	30011	0	0.00	22615
total_ic_mou_6	float64	30011	0	0.00	20602
total_ic_mou_7	float64	30011	0	0.00	20711
total_ic_mou_8	float64	30011	0	0.00	20096
total_ic_mou_9	float64	30011	0	0.00	19437
total_rech_num_6	int64	30011	0	0.00	102
total_rech_num_7	int64	30011	0	0.00	101
total_rech_num_8	int64	30011	0	0.00	96
total_rech_num_9	int64	30011	0	0.00	96
total_rech_amt_6	int64	30011	0	0.00	2241
total_rech_amt_7	int64	30011	0	0.00	2265
total_rech_amt_8	int64	30011	0	0.00	2299
total_rech_amt_9	int64	30011	0	0.00	2248
max_rech_amt_6	int64	30011	0	0.00	170
max_rech_amt_7	int64	30011	0	0.00	151
max_rech_amt_8	int64	30011	0	0.00	182
max_rech_amt_9	int64	30011	0	0.00	186
last_day_rch_amt_6	int64	30011	0	0.00	158
last_day_rch_amt_7	int64	30011	0	0.00	149
Average_rech_amt_6n7	float64	30011	0	0.00	3025

	total_rech_num_6	total_rech_amt_6	max_rech_amt_6	date_of_last_rech_6
mobile_number
7001588448	0	0	0	NaT
7001223277	0	0	0	NaT
7000721536	0	0	0	NaT
7001490351	0	0	0	NaT
7000665415	0	0	0	NaT

	total_rech_num_7	total_rech_amt_7	max_rech_amt_7	date_of_last_rech_7	Average_rech_amt_6n7
mobile_number
7000369789	0	0	0	NaT	393.0
7001967148	0	0	0	NaT	500.5
7000066601	0	0	0	NaT	490.0
7001189556	0	0	0	NaT	523.5
7002024450	0	0	0	NaT	493.0

	total_rech_num_8	total_rech_amt_8	max_rech_amt_8	date_of_last_rech_8
mobile_number
7000340381	0	0	0	NaT
7000608224	0	0	0	NaT
7000369789	0	0	0	NaT
7000248548	0	0	0	NaT
7001967063	0	0	0	NaT

	total_rech_num_9	total_rech_amt_9	max_rech_amt_9	date_of_last_rech_9
mobile_number
7000340381	0	0	0	NaT
7000854899	0	0	0	NaT
7000369789	0	0	0	NaT
7001967063	0	0	0	NaT
7000066601	0	0	0	NaT

	arpu_6	arpu_7	arpu_8	onnet_mou_6	onnet_mou_7	onnet_mou_8	offnet_mou_6	offnet_mou_7	offnet_mou_8	roam_ic_mou_6	roam_ic_mou_7	roam_ic_mou_8	roam_og_mou_6	roam_og_mou_7	roam_og_mou_8	loc_og_t2t_mou_6	loc_og_t2t_mou_7	loc_og_t2t_mou_8	loc_og_t2m_mou_6	loc_og_t2m_mou_7	loc_og_t2m_mou_8	loc_og_t2f_mou_6	loc_og_t2f_mou_7	loc_og_t2f_mou_8	loc_og_t2c_mou_6	loc_og_t2c_mou_7	loc_og_t2c_mou_8	loc_og_mou_6	loc_og_mou_7	loc_og_mou_8	std_og_t2t_mou_6	std_og_t2t_mou_7	std_og_t2t_mou_8	std_og_t2m_mou_6	std_og_t2m_mou_7	std_og_t2m_mou_8	std_og_t2f_mou_6	std_og_t2f_mou_7	std_og_t2f_mou_8	std_og_mou_6	std_og_mou_7	std_og_mou_8	isd_og_mou_6	isd_og_mou_7	isd_og_mou_8	spl_og_mou_6	spl_og_mou_7	spl_og_mou_8	og_others_6	og_others_7	og_others_8	total_og_mou_6	total_og_mou_7	total_og_mou_8	loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	loc_ic_t2t_mou_8	loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	loc_ic_mou_6	loc_ic_mou_7	loc_ic_mou_8	std_ic_t2t_mou_6	std_ic_t2t_mou_7	std_ic_t2t_mou_8	std_ic_t2m_mou_6	std_ic_t2m_mou_7	std_ic_t2m_mou_8	std_ic_t2f_mou_6	std_ic_t2f_mou_7	std_ic_t2f_mou_8	std_ic_mou_6	std_ic_mou_7	std_ic_mou_8	total_ic_mou_6	total_ic_mou_7	total_ic_mou_8	spl_ic_mou_6	spl_ic_mou_7	spl_ic_mou_8	isd_ic_mou_6	isd_ic_mou_7	isd_ic_mou_8	ic_others_6	ic_others_7	ic_others_8	total_rech_num_6	total_rech_num_7	total_rech_num_8	total_rech_amt_6	total_rech_amt_7	total_rech_amt_8	max_rech_amt_6	max_rech_amt_7	max_rech_amt_8	last_day_rch_amt_6	last_day_rch_amt_7	last_day_rch_amt_8	vol_2g_mb_6	vol_2g_mb_7	vol_2g_mb_8	vol_3g_mb_6	vol_3g_mb_7	vol_3g_mb_8	monthly_2g_6	monthly_2g_7	monthly_2g_8	sachet_2g_6	sachet_2g_7	sachet_2g_8	monthly_3g_6	monthly_3g_7	monthly_3g_8	sachet_3g_6	sachet_3g_7	sachet_3g_8	aon	vbc_3g_8	vbc_3g_7	vbc_3g_6	Average_rech_amt_6n7	Churn
mobile_number
7000701601	1069.180	1349.850	3171.480	57.84	54.68	52.29	453.43	567.16	325.91	16.23	33.49	31.64	23.74	12.59	38.06	51.39	31.38	40.28	308.63	447.38	162.28	62.13	55.14	53.23	0.0	0.0	0.00	422.16	533.91	255.79	4.30	23.29	12.01	49.89	31.76	49.14	6.66	20.08	16.68	60.86	75.14	77.84	0.0	0.18	10.01	4.50	0.00	6.50	0.00	0.0	0.0	487.53	609.24	350.16	58.14	32.26	27.31	217.56	221.49	121.19	152.16	101.46	39.53	427.88	355.23	188.04	36.89	11.83	30.39	91.44	126.99	141.33	52.19	34.24	22.21	180.54	173.08	193.94	626.46	558.04	428.74	0.21	0.0	0.0	2.06	14.53	31.59	15.74	15.19	15.14	5	5	7	1580	790	3638	1580	790	1580	0	0	779	0.0	0.0	0.00	0.0	0.00	0.00	0	0	0	0	0	0	0	0	0	0	0	0	802	57.74	19.38	18.74	1185.0	1
7001524846	378.721	492.223	137.362	413.69	351.03	35.08	94.66	80.63	136.48	0.00	0.00	0.00	0.00	0.00	0.00	297.13	217.59	12.49	80.96	70.58	50.54	0.00	0.00	0.00	0.0	0.0	7.15	378.09	288.18	63.04	116.56	133.43	22.58	13.69	10.04	75.69	0.00	0.00	0.00	130.26	143.48	98.28	0.0	0.00	0.00	0.00	0.00	10.23	0.00	0.0	0.0	508.36	431.66	171.56	23.84	9.84	0.31	57.58	13.98	15.48	0.00	0.00	0.00	81.43	23.83	15.79	0.00	0.58	0.10	22.43	4.08	0.65	0.00	0.00	0.00	22.43	4.66	0.75	103.86	28.49	16.54	0.00	0.0	0.0	0.00	0.00	0.00	0.00	0.00	0.00	19	21	14	437	601	120	90	154	30	50	0	10	0.0	356.0	0.03	0.0	750.95	11.94	0	1	0	0	1	3	0	0	0	0	0	0	315	21.03	910.65	122.16	519.0	0
7002191713	492.846	205.671	593.260	501.76	108.39	534.24	413.31	119.28	482.46	23.53	144.24	72.11	7.98	35.26	1.44	49.63	6.19	36.01	151.13	47.28	294.46	4.54	0.00	23.51	0.0	0.0	0.49	205.31	53.48	353.99	446.41	85.98	498.23	255.36	52.94	156.94	0.00	0.00	0.00	701.78	138.93	655.18	0.0	0.00	1.29	0.00	0.00	4.78	0.00	0.0	0.0	907.09	192.41	1015.26	67.88	7.58	52.58	142.88	18.53	195.18	4.81	0.00	7.49	215.58	26.11	255.26	115.68	38.29	154.58	308.13	29.79	317.91	0.00	0.00	1.91	423.81	68.09	474.41	968.61	172.58	1144.53	0.45	0.0	0.0	245.28	62.11	393.39	83.48	16.24	21.44	6	4	11	507	253	717	110	110	130	110	50	0	0.0	0.0	0.02	0.0	0.00	0.00	0	0	0	0	0	3	0	0	0	0	0	0	2607	0.00	0.00	0.00	380.0	0
7000875565	430.975	299.869	187.894	50.51	74.01	70.61	296.29	229.74	162.76	0.00	2.83	0.00	0.00	17.74	0.00	42.61	65.16	67.38	273.29	145.99	128.28	0.00	4.48	10.26	0.0	0.0	0.00	315.91	215.64	205.93	7.89	2.58	3.23	22.99	64.51	18.29	0.00	0.00	0.00	30.89	67.09	21.53	0.0	0.00	0.00	0.00	3.26	5.91	0.00	0.0	0.0	346.81	286.01	233.38	41.33	71.44	28.89	226.81	149.69	150.16	8.71	8.68	32.71	276.86	229.83	211.78	68.79	78.64	6.33	18.68	73.08	73.93	0.51	0.00	2.18	87.99	151.73	82.44	364.86	381.56	294.46	0.00	0.0	0.0	0.00	0.00	0.23	0.00	0.00	0.00	10	6	2	570	348	160	110	110	130	100	100	130	0.0	0.0	0.00	0.0	0.00	0.00	0	0	0	0	0	0	0	0	0	0	0	0	511	0.00	2.45	21.89	459.0	0
7000187447	690.008	18.980	25.499	1185.91	9.28	7.79	61.64	0.00	5.54	0.00	4.76	4.81	0.00	8.46	13.34	38.99	0.00	0.00	58.54	0.00	0.00	0.00	0.00	0.00	0.0	0.0	0.00	97.54	0.00	0.00	1146.91	0.81	0.00	1.55	0.00	0.00	0.00	0.00	0.00	1148.46	0.81	0.00	0.0	0.00	0.00	2.58	0.00	0.00	0.93	0.0	0.0	1249.53	0.81	0.00	34.54	0.00	0.00	47.41	2.31	0.00	0.00	0.00	0.00	81.96	2.31	0.00	8.63	0.00	0.00	1.28	0.00	0.00	0.00	0.00	0.00	9.91	0.00	0.00	91.88	2.31	0.00	0.00	0.0	0.0	0.00	0.00	0.00	0.00	0.00	0.00	19	2	4	816	0	30	110	0	30	30	0	0	0.0	0.0	0.00	0.0	0.00	0.00	0	0	0	0	0	0	0	0	0	0	0	0	667	0.00	0.00	0.00	408.0	0

VAR1	VAR2	CORR
isd_og_mou_8	isd_og_mou_7	0.96
isd_og_mou_8	isd_og_mou_6	0.95
arpu_8	total_rech_amt_8	0.95
isd_og_mou_7	isd_og_mou_6	0.95
arpu_6	total_rech_amt_6	0.94
total_rech_amt_7	arpu_7	0.94
Average_rech_amt_6n7	arpu_7	0.91
total_rech_amt_7	Average_rech_amt_6n7	0.91
total_ic_mou_6	loc_ic_mou_6	0.90
Average_rech_amt_6n7	total_rech_amt_6	0.90
arpu_6	Average_rech_amt_6n7	0.89
loc_ic_mou_8	total_ic_mou_8	0.89
loc_ic_mou_7	total_ic_mou_7	0.88
loc_ic_mou_7	loc_ic_mou_8	0.85
std_og_t2t_mou_8	onnet_mou_8	0.85
loc_ic_mou_8	loc_ic_t2m_mou_8	0.85
loc_ic_t2m_mou_6	loc_ic_mou_6	0.85
std_og_t2m_mou_8	offnet_mou_8	0.85
std_og_t2t_mou_7	onnet_mou_7	0.84
std_og_mou_8	total_og_mou_8	0.84
loc_og_mou_8	loc_og_mou_7	0.84
std_ic_mou_8	std_ic_t2m_mou_8	0.84
std_og_t2t_mou_6	onnet_mou_6	0.84
offnet_mou_7	std_og_t2m_mou_7	0.84
total_og_mou_7	std_og_mou_7	0.83
loc_ic_mou_7	loc_ic_mou_6	0.83
total_ic_mou_7	total_ic_mou_8	0.83
loc_og_t2t_mou_8	loc_og_t2t_mou_7	0.83
loc_ic_mou_7	loc_ic_t2m_mou_7	0.83
loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	0.82
loc_og_t2f_mou_8	loc_og_t2f_mou_7	0.82
loc_og_t2m_mou_8	loc_og_t2m_mou_7	0.82
onnet_mou_7	onnet_mou_8	0.82
std_ic_t2m_mou_6	std_ic_mou_6	0.82
std_og_t2t_mou_7	std_og_t2t_mou_8	0.82
std_ic_t2m_mou_7	std_ic_mou_7	0.81
loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	0.81
std_og_mou_8	std_og_mou_7	0.81
offnet_mou_6	std_og_t2m_mou_6	0.81
total_ic_mou_6	total_ic_mou_7	0.81
loc_ic_t2t_mou_8	loc_ic_t2t_mou_7	0.81
total_og_mou_6	std_og_mou_6	0.80
loc_og_mou_6	loc_og_mou_7	0.80
loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	0.80
loc_og_t2t_mou_6	loc_og_t2t_mou_7	0.80
std_og_t2m_mou_8	std_og_t2m_mou_7	0.79
loc_og_t2f_mou_7	loc_og_t2f_mou_6	0.79
loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	0.79
loc_og_mou_6	loc_og_t2m_mou_6	0.79
total_rech_num_8	total_rech_num_7	0.78
loc_og_t2m_mou_7	loc_og_t2m_mou_6	0.78
arpu_8	Average_rech_amt_6n7	0.78
offnet_mou_7	offnet_mou_8	0.78
loc_og_t2t_mou_8	loc_og_mou_8	0.77
arpu_7	total_rech_amt_8	0.77
arpu_8	arpu_7	0.77
arpu_8	total_rech_amt_7	0.77
loc_og_t2m_mou_8	loc_og_mou_8	0.77
total_og_mou_7	total_og_mou_8	0.77
std_og_t2f_mou_7	std_og_t2f_mou_8	0.77
loc_og_mou_7	loc_og_t2t_mou_7	0.77
total_ic_mou_8	loc_ic_t2m_mou_8	0.76
std_ic_mou_8	std_ic_mou_7	0.76
loc_ic_t2m_mou_6	total_ic_mou_6	0.76
isd_ic_mou_7	isd_ic_mou_6	0.75
isd_ic_mou_8	isd_ic_mou_7	0.75
loc_og_mou_6	loc_og_t2t_mou_6	0.75
std_ic_mou_6	std_ic_mou_7	0.75
loc_ic_mou_7	total_ic_mou_8	0.75
loc_og_t2m_mou_7	loc_og_mou_7	0.75
loc_ic_mou_8	loc_ic_mou_6	0.75
total_ic_mou_7	loc_ic_mou_8	0.75
Average_rech_amt_6n7	total_rech_amt_8	0.75
loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	0.75
vol_3g_mb_8	vol_3g_mb_7	0.75
std_og_t2m_mou_6	std_og_t2m_mou_7	0.75
std_og_mou_8	std_og_t2m_mou_8	0.75
std_og_t2m_mou_6	std_og_mou_6	0.74
std_og_mou_8	std_og_t2t_mou_8	0.74
std_ic_t2f_mou_7	std_ic_t2f_mou_6	0.74
std_og_t2m_mou_7	std_og_mou_7	0.74
loc_ic_mou_7	total_ic_mou_6	0.74
loc_ic_t2m_mou_7	total_ic_mou_7	0.74
std_og_mou_6	std_og_t2t_mou_6	0.74
std_ic_t2t_mou_6	std_ic_t2t_mou_7	0.74
std_ic_t2t_mou_8	std_ic_t2t_mou_7	0.73
loc_og_t2f_mou_8	loc_og_t2f_mou_6	0.73
std_og_t2t_mou_7	std_og_t2t_mou_6	0.73
std_ic_t2m_mou_7	std_ic_t2m_mou_8	0.73
onnet_mou_7	onnet_mou_6	0.73
total_rech_amt_7	total_rech_amt_8	0.73
loc_og_mou_6	loc_og_mou_8	0.73
total_ic_mou_6	total_ic_mou_8	0.73
std_og_t2t_mou_7	std_og_mou_7	0.73
std_og_mou_6	std_og_mou_7	0.73
total_ic_mou_7	loc_ic_mou_6	0.72
std_ic_t2f_mou_7	std_ic_t2f_mou_8	0.72
offnet_mou_6	offnet_mou_7	0.72
loc_ic_t2m_mou_6	loc_ic_t2m_mou_8	0.72
loc_ic_t2m_mou_7	loc_ic_mou_8	0.72
total_og_mou_8	offnet_mou_8	0.72
std_ic_t2m_mou_7	std_ic_t2m_mou_6	0.72
loc_og_t2t_mou_8	loc_og_t2t_mou_6	0.72
total_og_mou_8	onnet_mou_8	0.71
vbc_3g_8	vbc_3g_7	0.71
vol_3g_mb_6	vol_3g_mb_7	0.71
std_og_t2f_mou_6	std_og_t2f_mou_7	0.71
total_og_mou_7	onnet_mou_7	0.71
ic_others_8	ic_others_7	0.71
total_og_mou_6	offnet_mou_6	0.70
total_rech_amt_6	arpu_7	0.70
total_og_mou_7	offnet_mou_7	0.70
std_ic_t2t_mou_7	std_ic_mou_7	0.70
arpu_6	arpu_7	0.70
total_og_mou_6	onnet_mou_6	0.70
loc_ic_t2t_mou_6	loc_ic_t2t_mou_8	0.70
loc_ic_mou_7	loc_ic_t2m_mou_8	0.70
last_day_rch_amt_8	max_rech_amt_8	0.69
std_og_t2t_mou_7	onnet_mou_8	0.69
vbc_3g_7	vbc_3g_6	0.69
loc_og_t2m_mou_8	loc_og_t2m_mou_6	0.69
loc_ic_t2m_mou_7	loc_ic_mou_6	0.69
total_rech_num_6	total_rech_num_7	0.69
vol_2g_mb_7	vol_2g_mb_8	0.69
std_og_t2t_mou_8	onnet_mou_7	0.69
loc_ic_mou_7	loc_ic_t2m_mou_6	0.68
arpu_6	total_rech_amt_7	0.68
loc_ic_mou_7	loc_ic_t2t_mou_7	0.68
total_ic_mou_6	loc_ic_mou_8	0.68
std_ic_t2f_mou_8	std_ic_t2f_mou_6	0.67
ic_others_7	ic_others_6	0.67
loc_ic_t2t_mou_6	loc_ic_mou_6	0.67
loc_ic_t2t_mou_8	loc_ic_mou_8	0.67
vol_2g_mb_7	vol_2g_mb_6	0.67
loc_ic_t2f_mou_6	loc_ic_t2f_mou_8	0.67
total_og_mou_6	total_og_mou_7	0.67
vol_3g_mb_8	vol_3g_mb_6	0.67
std_ic_t2t_mou_6	std_ic_mou_6	0.67
total_ic_mou_8	loc_ic_mou_6	0.66
std_ic_t2t_mou_8	std_ic_mou_8	0.66
total_og_mou_7	std_og_mou_8	0.66
std_og_t2m_mou_8	offnet_mou_7	0.66
std_ic_mou_8	std_ic_mou_6	0.66
vbc_3g_7	vol_3g_mb_7	0.65
max_rech_amt_6	last_day_rch_amt_6	0.65
std_og_t2m_mou_7	offnet_mou_8	0.65
std_og_t2f_mou_6	std_og_t2f_mou_8	0.65
loc_og_t2t_mou_8	loc_og_mou_7	0.65
arpu_6	total_rech_amt_8	0.64
arpu_6	arpu_8	0.64
total_og_mou_8	std_og_mou_7	0.64
std_og_t2m_mou_8	total_og_mou_8	0.64
loc_ic_t2m_mou_6	loc_ic_mou_8	0.64
roam_og_mou_6	roam_ic_mou_6	0.64
std_ic_t2m_mou_7	std_ic_mou_8	0.64
loc_og_mou_8	loc_og_t2t_mou_7	0.64
loc_ic_t2m_mou_7	total_ic_mou_8	0.64
total_rech_amt_7	total_rech_amt_6	0.64
total_ic_mou_7	loc_ic_t2m_mou_8	0.63
vol_3g_mb_6	vbc_3g_6	0.63
vbc_3g_8	vol_3g_mb_8	0.63
total_ic_mou_6	loc_ic_t2m_mou_7	0.63
roam_ic_mou_7	roam_og_mou_7	0.63
onnet_mou_8	onnet_mou_6	0.63
loc_og_t2m_mou_8	loc_og_mou_7	0.63
arpu_8	total_rech_amt_6	0.63
std_og_t2t_mou_8	std_og_t2t_mou_6	0.63
loc_ic_t2m_mou_8	loc_ic_mou_6	0.63
loc_og_t2t_mou_6	loc_og_mou_7	0.63
total_og_mou_7	std_og_t2m_mou_7	0.63
ic_others_8	ic_others_6	0.63
std_ic_t2m_mou_7	std_ic_mou_6	0.63
loc_og_mou_8	loc_og_t2m_mou_7	0.63
std_ic_t2m_mou_6	std_ic_t2m_mou_8	0.63
total_rech_amt_6	total_rech_amt_8	0.63
isd_ic_mou_8	isd_ic_mou_6	0.62
std_og_t2t_mou_8	std_og_mou_7	0.62
total_og_mou_8	std_og_t2t_mou_8	0.61
loc_og_t2m_mou_6	loc_og_mou_7	0.61
onnet_mou_7	std_og_t2t_mou_6	0.61
vbc_3g_8	vbc_3g_6	0.61
loc_og_mou_6	loc_og_t2m_mou_7	0.61
std_og_mou_8	onnet_mou_8	0.61
std_ic_t2m_mou_8	std_ic_mou_7	0.61
last_day_rch_amt_7	max_rech_amt_7	0.61
std_og_t2m_mou_6	offnet_mou_7	0.61
std_og_mou_8	std_og_mou_6	0.61
max_rech_amt_6	max_rech_amt_8	0.60
std_og_mou_8	std_og_t2m_mou_7	0.60
total_rech_num_8	total_rech_num_6	0.60
total_ic_mou_7	loc_ic_t2t_mou_7	0.60
loc_ic_t2m_mou_6	total_ic_mou_7	0.60
roam_og_mou_8	roam_ic_mou_8	0.60
std_og_t2t_mou_7	onnet_mou_6	0.60
total_og_mou_7	std_og_t2t_mou_7	0.60
std_og_mou_8	std_og_t2t_mou_7	0.60
loc_og_mou_6	loc_og_t2t_mou_7	0.60
total_og_mou_6	std_og_t2m_mou_6	0.60
std_og_mou_8	offnet_mou_8	0.60
std_og_t2m_mou_8	std_og_t2m_mou_6	0.60
std_og_mou_6	onnet_mou_6	0.59
loc_ic_t2t_mou_8	total_ic_mou_8	0.59
onnet_mou_7	std_og_mou_7	0.59
total_ic_mou_6	loc_ic_t2t_mou_6	0.59
std_og_t2m_mou_8	std_og_mou_7	0.59
offnet_mou_6	offnet_mou_8	0.59
std_ic_t2t_mou_6	std_ic_t2t_mou_8	0.59
loc_ic_mou_7	loc_ic_t2t_mou_8	0.59
total_og_mou_7	onnet_mou_8	0.58
std_ic_t2m_mou_6	std_ic_mou_7	0.58
total_og_mou_6	std_og_t2t_mou_6	0.58
offnet_mou_6	std_og_t2m_mou_7	0.58
roam_og_mou_8	roam_og_mou_7	0.58
total_ic_mou_6	loc_ic_t2m_mou_8	0.58
spl_og_mou_7	spl_og_mou_8	0.57
total_og_mou_7	std_og_mou_6	0.57
offnet_mou_7	std_og_mou_7	0.57
loc_ic_mou_7	loc_ic_t2t_mou_6	0.57
loc_og_t2t_mou_8	loc_og_mou_6	0.57
std_og_t2m_mou_6	std_og_mou_7	0.56
max_rech_amt_7	max_rech_amt_8	0.56
loc_ic_t2m_mou_6	total_ic_mou_8	0.56
spl_og_mou_7	spl_og_mou_6	0.56
roam_ic_mou_7	roam_ic_mou_8	0.56
std_ic_t2m_mou_8	std_ic_mou_6	0.56
loc_og_mou_8	loc_og_t2t_mou_6	0.56
loc_ic_mou_8	loc_ic_t2t_mou_7	0.56
loc_og_mou_8	loc_og_t2m_mou_6	0.56
total_og_mou_8	onnet_mou_7	0.56
total_og_mou_6	total_og_mou_8	0.55
loc_og_t2m_mou_8	loc_og_mou_6	0.55
std_ic_t2t_mou_6	std_ic_mou_7	0.55
loc_ic_t2t_mou_7	loc_ic_mou_6	0.55
total_og_mou_7	offnet_mou_8	0.54
std_og_t2t_mou_7	std_og_mou_6	0.54
std_ic_mou_8	std_ic_t2m_mou_6	0.54
offnet_mou_6	std_og_mou_6	0.54
std_og_mou_6	std_og_t2m_mou_7	0.54
total_og_mou_8	offnet_mou_7	0.54
spl_og_mou_7	loc_og_t2c_mou_7	0.53
std_og_t2t_mou_6	std_og_mou_7	0.53
total_og_mou_6	std_og_mou_7	0.53
std_og_t2t_mou_6	onnet_mou_8	0.53
std_ic_t2t_mou_8	std_ic_mou_7	0.53
loc_og_t2c_mou_8	loc_og_t2c_mou_7	0.53
Average_rech_amt_6n7	isd_og_mou_7	0.53
std_og_t2t_mou_8	onnet_mou_6	0.52
vol_2g_mb_8	vol_2g_mb_6	0.52
isd_og_mou_7	arpu_7	0.52
loc_ic_t2t_mou_8	loc_ic_mou_6	0.51
arpu_8	total_og_mou_8	0.51
total_ic_mou_7	loc_ic_t2t_mou_8	0.51
vol_3g_mb_7	vbc_3g_6	0.51
vbc_3g_8	vol_3g_mb_7	0.51
total_og_mou_6	arpu_6	0.51
total_rech_amt_7	isd_og_mou_7	0.50
roam_og_mou_7	roam_og_mou_6	0.50
onnet_mou_8	std_og_mou_7	0.50
Average_rech_amt_6n7	isd_og_mou_6	0.50
loc_ic_t2t_mou_6	total_ic_mou_7	0.50
loc_ic_t2t_mou_6	loc_ic_mou_8	0.50
std_ic_t2t_mou_7	std_ic_mou_6	0.50
max_rech_amt_6	max_rech_amt_7	0.50
isd_og_mou_8	Average_rech_amt_6n7	0.50
std_ic_mou_8	std_ic_t2t_mou_7	0.50
loc_ic_t2m_mou_6	loc_og_t2m_mou_6	0.50
total_og_mou_6	total_rech_amt_6	0.49
loc_ic_t2t_mou_7	total_ic_mou_8	0.49
total_og_mou_7	std_og_t2m_mou_8	0.49
isd_og_mou_8	total_rech_amt_8	0.49
total_og_mou_7	std_og_t2t_mou_8	0.49
total_og_mou_8	total_rech_amt_8	0.49
loc_og_t2m_mou_7	loc_ic_t2m_mou_7	0.49
total_ic_mou_6	loc_ic_t2t_mou_7	0.49
vbc_3g_7	vol_3g_mb_8	0.49
std_og_mou_8	onnet_mou_7	0.49
loc_og_t2m_mou_8	loc_ic_t2m_mou_8	0.49
total_og_mou_7	onnet_mou_6	0.48
total_og_mou_7	offnet_mou_6	0.48
std_og_t2m_mou_6	offnet_mou_8	0.48
total_og_mou_7	arpu_7	0.48
max_rech_amt_8	total_rech_amt_8	0.48
isd_og_mou_8	arpu_7	0.48
total_og_mou_8	std_og_t2m_mou_7	0.48
arpu_6	isd_og_mou_6	0.48
total_og_mou_6	onnet_mou_7	0.48
isd_og_mou_7	total_rech_amt_8	0.48
isd_og_mou_8	arpu_8	0.48
loc_og_t2c_mou_6	spl_og_mou_6	0.48
offnet_mou_6	std_og_t2m_mou_8	0.47
vol_3g_mb_8	vbc_3g_6	0.47
total_rech_amt_6	isd_og_mou_6	0.47
onnet_mou_7	loc_og_t2t_mou_7	0.47
std_og_t2t_mou_8	std_og_mou_6	0.47
arpu_6	isd_og_mou_7	0.47
total_og_mou_6	offnet_mou_7	0.47
offnet_mou_6	loc_og_t2m_mou_6	0.47
std_og_t2t_mou_7	total_og_mou_8	0.47
loc_og_t2t_mou_6	onnet_mou_6	0.47
arpu_8	offnet_mou_8	0.47
roam_ic_mou_7	roam_ic_mou_6	0.46
arpu_7	isd_og_mou_6	0.46
offnet_mou_8	total_rech_amt_8	0.46
total_og_mou_7	total_rech_amt_7	0.46
spl_og_mou_8	loc_og_t2c_mou_8	0.46
isd_og_mou_8	arpu_6	0.46
isd_og_mou_7	total_rech_amt_6	0.46
std_og_mou_8	std_og_t2m_mou_6	0.46
arpu_8	isd_og_mou_7	0.46
std_ic_mou_8	total_ic_mou_8	0.46
total_rech_amt_7	max_rech_amt_7	0.46
vbc_3g_7	vol_3g_mb_6	0.46
total_og_mou_8	std_og_mou_6	0.46
loc_og_t2t_mou_8	onnet_mou_8	0.46
arpu_6	offnet_mou_6	0.46
isd_og_mou_8	total_rech_amt_7	0.46
offnet_mou_6	total_rech_amt_6	0.45
total_ic_mou_6	loc_ic_t2t_mou_8	0.45
total_ic_mou_7	std_ic_mou_7	0.45
std_og_mou_8	offnet_mou_7	0.45
isd_og_mou_8	total_rech_amt_6	0.45
total_og_mou_7	std_og_t2m_mou_6	0.45
std_og_mou_8	std_og_t2t_mou_6	0.45
loc_og_mou_6	loc_ic_mou_6	0.45
total_ic_mou_6	std_ic_mou_6	0.44
loc_og_t2m_mou_8	loc_ic_mou_8	0.44
loc_og_t2m_mou_8	offnet_mou_8	0.44
total_og_mou_6	loc_og_mou_6	0.44
total_og_mou_6	std_og_mou_8	0.44
isd_og_mou_6	total_rech_amt_8	0.44
std_og_mou_7	offnet_mou_8	0.44
std_og_t2m_mou_8	std_og_mou_6	0.44
loc_og_t2m_mou_6	loc_ic_mou_6	0.44
std_ic_t2t_mou_6	std_ic_mou_8	0.44
loc_og_mou_8	loc_ic_mou_8	0.44
offnet_mou_7	arpu_7	0.44
arpu_8	isd_og_mou_6	0.43
loc_og_t2m_mou_7	loc_ic_t2m_mou_8	0.43
max_rech_amt_6	total_rech_amt_6	0.43
loc_og_t2m_mou_8	loc_ic_t2m_mou_7	0.43
total_rech_amt_7	isd_og_mou_6	0.43
vbc_3g_8	vol_3g_mb_6	0.43
total_rech_amt_7	offnet_mou_7	0.43
loc_ic_t2t_mou_6	total_ic_mou_8	0.43
onnet_mou_7	std_og_mou_6	0.43
loc_og_mou_8	total_og_mou_8	0.42
loc_ic_t2m_mou_7	loc_og_t2m_mou_6	0.42
total_og_mou_6	Average_rech_amt_6n7	0.42
loc_ic_mou_7	loc_og_mou_7	0.42
loc_ic_mou_7	loc_og_t2m_mou_7	0.42
total_og_mou_7	Average_rech_amt_6n7	0.42
last_day_rch_amt_6	max_rech_amt_8	0.42
std_ic_t2t_mou_8	std_ic_mou_6	0.42
loc_ic_t2m_mou_6	loc_og_t2m_mou_7	0.42
loc_og_t2m_mou_7	offnet_mou_7	0.42
total_og_mou_6	onnet_mou_8	0.42
spl_og_mou_8	spl_og_mou_6	0.41
last_day_rch_amt_8	max_rech_amt_7	0.41
offnet_mou_6	Average_rech_amt_6n7	0.41
max_rech_amt_6	last_day_rch_amt_8	0.41
loc_ic_t2m_mou_6	loc_og_mou_6	0.41
total_og_mou_8	onnet_mou_6	0.41

VAR1	VAR2	CORR
og_others_7	og_others_8	1.00
arpu_8	total_rech_amt_8	0.96
arpu_6	total_rech_amt_6	0.95
std_og_mou_8	total_og_mou_8	0.95
total_rech_amt_7	arpu_7	0.95
std_og_t2t_mou_7	onnet_mou_7	0.95
total_og_mou_7	std_og_mou_7	0.94
og_others_8	loc_og_t2f_mou_6	0.93
std_og_t2t_mou_8	onnet_mou_8	0.93
loc_og_t2f_mou_7	loc_og_t2f_mou_6	0.93
og_others_7	loc_og_t2f_mou_6	0.93
total_og_mou_6	std_og_mou_6	0.92
offnet_mou_6	std_og_t2m_mou_6	0.92
offnet_mou_7	std_og_t2m_mou_7	0.92
std_og_t2t_mou_6	onnet_mou_6	0.92
std_ic_mou_8	std_ic_t2m_mou_8	0.92
loc_og_t2f_mou_7	og_others_8	0.91
loc_og_t2f_mou_7	og_others_7	0.91
loc_ic_mou_8	loc_ic_t2m_mou_8	0.90
loc_ic_t2m_mou_6	loc_ic_mou_6	0.90
loc_ic_mou_8	total_ic_mou_8	0.89
loc_og_t2m_mou_8	loc_og_mou_8	0.88
total_ic_mou_6	loc_ic_mou_6	0.87
std_og_t2m_mou_8	offnet_mou_8	0.87
loc_ic_mou_7	total_ic_mou_7	0.86
loc_ic_mou_7	loc_ic_t2m_mou_7	0.84
loc_og_t2m_mou_7	loc_og_mou_7	0.84
std_ic_t2m_mou_7	std_ic_mou_7	0.82
total_ic_mou_8	loc_ic_t2m_mou_8	0.81
std_og_mou_8	std_og_t2t_mou_8	0.79
std_ic_t2t_mou_6	std_ic_t2t_mou_7	0.78
Average_rech_amt_6n7	arpu_7	0.77
loc_og_mou_6	loc_og_t2m_mou_6	0.77
loc_ic_t2m_mou_6	total_ic_mou_6	0.77
std_ic_t2m_mou_6	std_ic_mou_6	0.77
total_rech_amt_7	Average_rech_amt_6n7	0.76
Average_rech_amt_6n7	total_rech_amt_6	0.76
loc_og_mou_6	loc_og_t2t_mou_6	0.75
total_og_mou_8	std_og_t2t_mou_8	0.75
std_og_t2m_mou_7	std_og_mou_7	0.74
std_og_mou_8	onnet_mou_8	0.74
total_og_mou_8	onnet_mou_8	0.74
arpu_6	Average_rech_amt_6n7	0.73
loc_og_t2t_mou_8	loc_og_t2t_mou_7	0.73
loc_ic_t2t_mou_8	loc_ic_mou_8	0.73
loc_ic_t2t_mou_6	loc_ic_mou_6	0.72
max_rech_amt_6	last_day_rch_amt_6	0.72
std_og_t2m_mou_6	std_og_mou_6	0.72
std_ic_t2t_mou_6	std_ic_mou_6	0.72
roam_ic_mou_7	roam_ic_mou_8	0.72
total_og_mou_7	offnet_mou_7	0.72
loc_ic_t2m_mou_7	total_ic_mou_7	0.72
std_og_mou_8	std_og_t2m_mou_8	0.71
total_og_mou_8	offnet_mou_8	0.70
last_day_rch_amt_8	max_rech_amt_8	0.70
loc_og_mou_7	loc_og_t2t_mou_7	0.69
std_og_t2t_mou_7	std_og_mou_7	0.69
total_og_mou_7	std_og_t2m_mou_7	0.69
loc_ic_mou_7	loc_ic_t2t_mou_7	0.69
loc_og_t2t_mou_8	loc_og_mou_8	0.68
total_og_mou_6	offnet_mou_6	0.68
std_og_mou_6	std_og_t2t_mou_6	0.68
std_og_t2m_mou_8	total_og_mou_8	0.68
max_rech_amt_8	total_rech_amt_8	0.68
spl_og_mou_7	loc_og_t2c_mou_7	0.68
loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	0.67
vol_3g_mb_8	vol_3g_mb_7	0.67
std_og_t2t_mou_7	std_og_t2t_mou_6	0.67
total_og_mou_6	std_og_t2m_mou_6	0.67
offnet_mou_7	std_og_mou_7	0.66
total_og_mou_7	onnet_mou_7	0.66
onnet_mou_7	std_og_mou_7	0.65
loc_og_mou_8	loc_ic_mou_8	0.65
std_ic_t2t_mou_7	std_ic_mou_7	0.65
loc_og_t2m_mou_8	loc_ic_t2m_mou_8	0.65
roam_og_mou_8	roam_og_mou_7	0.65
total_og_mou_6	onnet_mou_6	0.65
total_og_mou_7	std_og_t2t_mou_7	0.64
loc_ic_t2t_mou_8	total_ic_mou_8	0.64
onnet_mou_7	onnet_mou_6	0.64
loc_og_mou_8	loc_ic_t2m_mou_8	0.64
onnet_mou_7	std_og_t2t_mou_6	0.63
loc_og_mou_8	loc_og_mou_7	0.63
total_ic_mou_6	loc_ic_t2t_mou_6	0.63
offnet_mou_6	std_og_mou_6	0.63
roam_og_mou_6	roam_ic_mou_6	0.63
std_og_mou_8	offnet_mou_8	0.63
roam_ic_mou_7	roam_ic_mou_6	0.62
arpu_8	max_rech_amt_8	0.62
std_og_t2m_mou_6	std_og_t2m_mou_7	0.62
total_og_mou_6	std_og_t2t_mou_6	0.62
vol_3g_mb_6	vbc_3g_6	0.62
onnet_mou_7	onnet_mou_8	0.62
loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	0.62
loc_og_mou_8	total_ic_mou_8	0.61
vbc_3g_8	vbc_3g_7	0.61
loc_og_t2m_mou_7	loc_og_t2m_mou_6	0.61
std_og_t2t_mou_7	onnet_mou_6	0.61
roam_og_mou_7	roam_og_mou_6	0.61
std_og_t2t_mou_7	std_og_t2t_mou_8	0.61
std_og_mou_6	onnet_mou_6	0.61
std_og_t2t_mou_8	onnet_mou_7	0.60
std_ic_mou_6	std_ic_mou_7	0.60
loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	0.60
arpu_8	total_og_mou_8	0.60
std_og_t2f_mou_7	std_og_t2f_mou_8	0.60
isd_og_mou_8	isd_og_mou_7	0.60
loc_og_mou_6	loc_og_mou_7	0.59
loc_og_t2m_mou_8	loc_og_t2m_mou_7	0.59
last_day_rch_amt_7	max_rech_amt_7	0.59
arpu_8	offnet_mou_8	0.59
std_og_mou_8	std_og_mou_7	0.58
total_og_mou_8	total_rech_amt_8	0.58
loc_og_t2m_mou_7	loc_ic_t2m_mou_7	0.58
loc_og_t2m_mou_8	loc_ic_mou_8	0.58
loc_ic_mou_7	loc_ic_mou_8	0.58
std_og_t2m_mou_8	std_og_t2m_mou_7	0.58
total_ic_mou_7	loc_ic_t2t_mou_7	0.58
offnet_mou_8	total_rech_amt_8	0.58
std_og_mou_6	std_og_mou_7	0.58
spl_og_mou_8	loc_og_t2c_mou_8	0.57
loc_ic_mou_7	loc_ic_mou_6	0.57
isd_ic_mou_7	isd_ic_mou_6	0.57
offnet_mou_6	offnet_mou_7	0.57
offnet_mou_7	offnet_mou_8	0.57
vol_3g_mb_6	vol_3g_mb_7	0.57
isd_og_mou_7	arpu_7	0.57
loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	0.57
total_rech_num_8	total_og_mou_8	0.57
loc_og_t2t_mou_6	loc_og_t2t_mou_7	0.56
std_og_t2t_mou_7	onnet_mou_8	0.56
total_og_mou_7	total_og_mou_8	0.56
vbc_3g_7	vol_3g_mb_7	0.56
total_rech_amt_7	isd_og_mou_7	0.56
std_ic_mou_8	total_ic_mou_8	0.56
loc_ic_t2t_mou_8	loc_ic_t2t_mou_7	0.56
loc_og_t2c_mou_6	spl_og_mou_6	0.56
std_og_t2m_mou_6	offnet_mou_7	0.55
ic_others_7	ic_others_6	0.55
total_og_mou_7	std_og_mou_8	0.55
total_ic_mou_6	total_ic_mou_7	0.55
total_rech_num_8	total_rech_amt_8	0.55
arpu_8	total_rech_num_8	0.54
loc_ic_t2m_mou_7	loc_ic_mou_6	0.54
std_ic_t2t_mou_8	std_ic_mou_8	0.54
loc_og_t2c_mou_6	loc_og_t2c_mou_7	0.54
std_og_t2m_mou_8	offnet_mou_7	0.54
total_rech_num_8	total_rech_num_7	0.54
total_ic_mou_7	std_ic_mou_7	0.54
std_ic_t2t_mou_7	std_ic_mou_6	0.54
loc_ic_mou_7	loc_ic_t2m_mou_6	0.54
offnet_mou_6	std_og_t2m_mou_7	0.54
std_og_mou_8	total_rech_num_8	0.54
loc_og_t2m_mou_8	total_ic_mou_8	0.54
total_ic_mou_7	total_ic_mou_8	0.54
std_ic_t2t_mou_8	std_ic_t2t_mou_7	0.54
total_ic_mou_6	std_ic_mou_6	0.53
vol_2g_mb_7	vol_2g_mb_6	0.53
vbc_3g_7	vbc_3g_6	0.53
arpu_6	isd_og_mou_6	0.53
total_og_mou_8	std_og_mou_7	0.52
std_ic_mou_8	std_ic_mou_7	0.52
total_rech_amt_6	isd_og_mou_6	0.52
loc_og_mou_8	loc_og_t2m_mou_7	0.52
arpu_8	std_og_mou_8	0.51
loc_og_t2m_mou_8	loc_og_mou_7	0.51
loc_ic_t2m_mou_7	loc_og_mou_7	0.51
loc_og_t2m_mou_6	loc_og_mou_7	0.51
arpu_8	total_rech_amt_7	0.51
total_og_mou_7	arpu_7	0.51
total_og_mou_6	total_og_mou_7	0.51
roam_ic_mou_7	roam_og_mou_7	0.51
loc_ic_t2m_mou_7	loc_ic_mou_8	0.51
total_og_mou_7	std_og_mou_6	0.51
loc_ic_t2m_mou_6	loc_og_t2m_mou_6	0.50
std_ic_t2m_mou_7	std_ic_t2m_mou_8	0.50
std_ic_t2m_mou_7	std_ic_t2m_mou_6	0.50
total_og_mou_6	std_og_mou_7	0.50
total_rech_amt_7	max_rech_amt_7	0.50
std_og_t2m_mou_7	offnet_mou_8	0.50
arpu_8	onnet_mou_8	0.50
onnet_mou_8	total_rech_amt_8	0.50
loc_ic_mou_7	loc_og_mou_7	0.50
total_ic_mou_7	loc_ic_mou_8	0.50
std_og_mou_8	total_rech_amt_8	0.50
loc_ic_mou_7	loc_ic_t2m_mou_8	0.50
last_day_rch_amt_8	total_rech_amt_8	0.50
std_og_t2f_mou_7	loc_og_t2f_mou_7	0.50
loc_og_t2t_mou_8	loc_og_mou_7	0.50
arpu_8	arpu_7	0.50
loc_ic_mou_7	total_ic_mou_6	0.49
std_ic_t2t_mou_6	std_ic_t2t_mou_8	0.49
loc_ic_mou_7	loc_og_t2m_mou_7	0.49
loc_ic_mou_7	total_ic_mou_8	0.49
vol_2g_mb_7	vol_2g_mb_8	0.49
vbc_3g_8	vol_3g_mb_8	0.49
loc_ic_mou_8	loc_ic_t2f_mou_8	0.49
arpu_7	total_rech_amt_8	0.48
std_og_t2f_mou_7	og_others_8	0.48
loc_og_t2t_mou_8	loc_ic_t2t_mou_8	0.48
vol_3g_mb_8	vol_3g_mb_6	0.48
std_ic_t2t_mou_6	std_ic_mou_7	0.48
isd_og_mou_7	isd_ic_mou_7	0.48
isd_ic_mou_8	isd_ic_mou_7	0.48
std_ic_t2m_mou_8	total_ic_mou_8	0.48
total_og_mou_7	total_rech_amt_7	0.48
std_og_t2f_mou_7	og_others_7	0.48
std_og_t2f_mou_7	loc_og_t2f_mou_6	0.48
loc_og_mou_8	total_og_mou_8	0.47
total_rech_num_8	onnet_mou_8	0.47
total_ic_mou_6	loc_ic_t2m_mou_7	0.47
loc_ic_mou_7	loc_ic_t2t_mou_8	0.47
total_og_mou_6	arpu_6	0.47
std_og_mou_8	std_og_t2t_mou_7	0.46
Average_rech_amt_6n7	isd_og_mou_7	0.46
spl_og_mou_7	spl_og_mou_8	0.46
loc_ic_t2f_mou_6	loc_ic_t2f_mou_8	0.46
roam_og_mou_8	last_day_rch_amt_8	0.46
total_og_mou_7	total_rech_num_7	0.46
total_og_mou_6	total_rech_amt_6	0.46
total_ic_mou_7	loc_ic_mou_6	0.46
std_ic_t2m_mou_7	std_ic_mou_8	0.46
loc_og_mou_8	loc_og_t2t_mou_7	0.46
max_rech_amt_6	total_rech_amt_6	0.46
arpu_8	roam_og_mou_8	0.45
loc_og_t2m_mou_6	loc_ic_mou_6	0.45
total_rech_num_8	std_og_t2t_mou_8	0.45
loc_og_mou_6	loc_og_t2t_mou_7	0.45
std_ic_t2m_mou_8	std_ic_mou_7	0.45
std_ic_t2m_mou_7	total_ic_mou_7	0.45
total_rech_num_6	total_rech_num_7	0.45
offnet_mou_7	arpu_7	0.45
loc_og_mou_6	loc_ic_mou_6	0.45
std_ic_t2f_mou_7	std_ic_t2f_mou_6	0.45
max_rech_amt_7	max_rech_amt_8	0.45
total_rech_amt_7	total_rech_amt_8	0.45
loc_og_mou_6	loc_og_t2m_mou_7	0.45
std_og_t2m_mou_8	std_og_mou_7	0.44
arpu_8	Average_rech_amt_6n7	0.44
total_ic_mou_7	loc_og_mou_7	0.44
std_og_mou_8	onnet_mou_7	0.44
loc_og_t2c_mou_8	loc_og_t2c_mou_7	0.44
roam_og_mou_8	roam_ic_mou_8	0.44
loc_ic_t2m_mou_6	total_ic_mou_7	0.44
roam_og_mou_8	total_rech_amt_8	0.44
arpu_8	last_day_rch_amt_8	0.44
ic_others_6	isd_ic_mou_6	0.44
loc_og_t2m_mou_8	offnet_mou_8	0.44
loc_ic_t2m_mou_7	total_ic_mou_8	0.43
std_og_t2t_mou_8	std_og_mou_7	0.43
loc_og_mou_8	offnet_mou_8	0.43
total_og_mou_6	total_rech_num_6	0.43
total_og_mou_8	onnet_mou_7	0.43
std_ic_t2m_mou_6	total_ic_mou_6	0.43
total_rech_num_8	offnet_mou_8	0.43
spl_og_mou_7	spl_og_mou_6	0.43
total_ic_mou_8	loc_ic_t2f_mou_8	0.43
std_ic_t2f_mou_7	std_og_t2f_mou_7	0.43
loc_og_t2t_mou_8	loc_ic_mou_8	0.42
loc_ic_t2m_mou_6	loc_og_mou_6	0.42
loc_ic_mou_7	loc_ic_t2f_mou_7	0.42
total_ic_mou_7	loc_ic_t2m_mou_8	0.42
max_rech_amt_6	max_rech_amt_8	0.42
loc_og_mou_8	loc_ic_t2t_mou_8	0.42
std_og_t2t_mou_7	std_og_mou_6	0.42
std_ic_t2m_mou_6	std_ic_mou_7	0.42
arpu_8	total_ic_mou_8	0.42
loc_og_t2m_mou_8	loc_ic_t2m_mou_7	0.42
last_day_rch_amt_7	max_rech_amt_8	0.42
arpu_6	offnet_mou_6	0.42
total_rech_amt_7	offnet_mou_7	0.42
loc_og_t2m_mou_7	total_ic_mou_7	0.42
total_og_mou_7	std_og_t2m_mou_8	0.42
Average_rech_amt_6n7	total_rech_amt_8	0.42
std_og_mou_7	arpu_7	0.41
std_og_mou_6	std_og_t2m_mou_7	0.41
total_rech_num_7	std_og_mou_7	0.41
total_og_mou_7	offnet_mou_8	0.41
spl_ic_mou_8	spl_ic_mou_6	0.41
std_og_t2t_mou_6	std_og_mou_7	0.41
offnet_mou_6	total_rech_amt_6	0.41
loc_og_t2t_mou_8	loc_og_t2t_mou_6	0.41
std_og_t2t_mou_7	total_og_mou_8	0.41
total_og_mou_8	total_ic_mou_8	0.41
std_og_t2m_mou_6	std_og_mou_7	0.41

	Churn
0	0.913598
1	0.086402

	onnet_mou_6	onnet_mou_7	onnet_mou_8	offnet_mou_6	offnet_mou_7	offnet_mou_8	roam_ic_mou_6	roam_ic_mou_7	roam_ic_mou_8	roam_og_mou_6	roam_og_mou_7	roam_og_mou_8	loc_og_t2t_mou_6	loc_og_t2t_mou_7	loc_og_t2t_mou_8	loc_og_t2m_mou_6	loc_og_t2m_mou_7	loc_og_t2m_mou_8	loc_og_t2f_mou_6	loc_og_t2f_mou_7	loc_og_t2f_mou_8	loc_og_t2c_mou_6	loc_og_t2c_mou_7	loc_og_t2c_mou_8	loc_og_mou_6	loc_og_mou_7	loc_og_mou_8	std_og_t2t_mou_6	std_og_t2t_mou_7	std_og_t2t_mou_8	std_og_t2m_mou_6	std_og_t2m_mou_7	std_og_t2m_mou_8	std_og_t2f_mou_6	std_og_t2f_mou_7	std_og_t2f_mou_8	std_og_mou_6	std_og_mou_7	std_og_mou_8	isd_og_mou_6	isd_og_mou_7	isd_og_mou_8	spl_og_mou_6	spl_og_mou_7	spl_og_mou_8	og_others_6	og_others_7	og_others_8	loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	loc_ic_t2t_mou_8	loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	loc_ic_mou_6	loc_ic_mou_7	loc_ic_mou_8	std_ic_t2t_mou_6	std_ic_t2t_mou_7	std_ic_t2t_mou_8	std_ic_t2m_mou_6	std_ic_t2m_mou_7	std_ic_t2m_mou_8	std_ic_t2f_mou_6	std_ic_t2f_mou_7	std_ic_t2f_mou_8	std_ic_mou_6	std_ic_mou_7	std_ic_mou_8	spl_ic_mou_6	spl_ic_mou_7	spl_ic_mou_8	isd_ic_mou_6	isd_ic_mou_7	isd_ic_mou_8	ic_others_6	ic_others_7	ic_others_8	total_rech_num_6	total_rech_num_7	total_rech_num_8	max_rech_amt_6	max_rech_amt_7	max_rech_amt_8	last_day_rch_amt_6	last_day_rch_amt_7	last_day_rch_amt_8	aon	Average_rech_amt_6n7	delta_vol_2g	delta_vol_3g	delta_total_og_mou	delta_total_ic_mou	delta_vbc_3g	delta_arpu	delta_total_rech_amt
0.9	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan
0.91	6.79	6.55	8.27	5.59	5.17	5.33	21.27	26.80	25.22	21.64	21.39	25.13	8.67	8.20	8.59	5.96	5.56	6.04	10.34	10.83	11.08	15.84	13.88	14.88	6.25	5.37	5.73	8.84	8.90	9.66	8.94	7.72	9.82	32.27	28.44	31.79	4.79	5.71	6.43	nan	nan	nan	10.11	9.09	9.16	12.39	nan	nan	7.05	7.29	7.90	6.00	6.19	6.01	11.11	10.60	10.81	6.05	5.48	5.59	10.03	10.84	11.79	9.58	8.84	10.15	19.50	19.89	19.54	8.69	8.96	9.10	7.14	nan	nan	22.19	19.35	22.84	24.14	22.01	22.76	4.35	4.35	4.76	9.43	10.00	14.68	0.00	0.00	11.11	2.25	3.40	33.68	33.56	9.39	9.84	36.51	8.39	8.31
0.92	7.08	7.25	8.36	6.69	5.49	6.01	22.72	27.49	26.73	19.34	21.90	24.03	9.73	8.85	8.65	6.41	5.68	6.11	11.24	11.91	11.09	16.45	13.57	14.71	5.67	6.21	6.26	9.33	8.79	9.32	8.36	9.37	9.79	28.52	32.50	32.02	6.12	5.68	6.76	inf	inf	nan	9.81	9.92	10.60	12.60	nan	nan	8.68	8.01	8.60	6.46	6.44	6.22	11.60	11.61	12.08	6.15	5.89	5.84	11.83	11.70	11.30	9.77	10.35	10.46	21.50	20.38	22.12	9.79	9.95	10.41	10.00	nan	inf	23.12	23.35	25.20	23.61	22.68	23.18	4.17	4.17	4.55	7.69	6.06	14.19	0.00	0.00	0.00	2.44	3.98	34.55	27.36	11.26	9.61	33.76	8.82	8.54
0.93	8.96	7.87	8.48	6.14	6.34	6.71	24.03	25.62	27.86	18.75	23.02	25.80	11.12	11.00	9.60	6.71	6.51	6.73	11.91	12.32	12.33	16.33	14.27	15.79	6.87	6.94	6.73	8.23	8.89	10.78	9.00	8.83	10.25	30.75	26.77	32.76	7.06	6.81	6.82	153.85	400.00	nan	10.76	10.46	10.50	12.94	nan	nan	8.59	8.54	8.44	6.99	6.62	6.99	12.21	13.49	12.40	6.68	6.84	6.50	12.40	13.98	14.13	11.02	10.25	12.01	23.85	22.24	23.09	10.83	11.75	12.50	9.09	nan	266.67	24.79	23.73	27.02	23.60	23.24	27.96	8.00	8.00	8.70	0.00	13.71	6.06	0.80	0.80	0.00	2.49	4.58	33.13	24.98	11.15	10.00	30.47	9.33	8.93
0.9400000000000001	7.71	7.97	9.40	6.70	7.47	7.79	22.41	32.61	27.41	23.62	26.30	24.59	11.66	12.58	10.94	7.67	7.62	7.34	15.38	14.43	14.39	16.09	18.10	15.52	7.80	7.97	7.60	9.38	9.02	11.82	10.12	9.63	11.32	27.40	29.92	31.63	7.21	7.01	8.00	90.91	104.00	inf	12.12	10.26	11.49	12.69	nan	nan	10.54	8.87	10.35	7.56	9.02	8.70	15.33	14.36	13.33	6.75	6.70	7.71	15.53	14.22	15.21	12.87	13.57	12.92	23.21	22.45	23.84	13.72	13.21	12.97	11.11	nan	90.91	27.04	22.91	22.11	27.73	27.17	25.21	3.70	3.70	4.00	14.29	14.32	13.71	0.00	0.00	0.80	1.69	4.83	36.42	28.15	12.40	12.12	30.69	10.38	9.47
0.9500000000000001	8.17	10.12	10.73	8.85	8.42	8.38	27.89	28.10	30.03	22.61	25.97	27.18	13.32	13.41	14.74	9.10	8.69	8.93	15.33	14.58	16.43	18.07	17.72	18.94	9.29	8.32	8.09	11.03	10.25	11.16	12.19	10.61	12.77	30.98	29.62	36.50	7.01	8.10	9.59	74.60	98.04	139.13	12.09	12.85	12.67	13.74	nan	nan	12.79	11.26	11.92	8.77	8.30	9.43	15.33	15.65	14.82	7.85	9.76	8.85	17.04	14.66	17.08	14.51	15.21	15.48	26.71	24.42	25.23	15.09	14.08	15.69	7.50	inf	19.05	29.73	26.89	27.12	29.18	29.49	31.88	7.14	7.14	7.69	25.00	9.89	14.32	0.00	8.73	0.00	2.32	6.70	33.06	21.43	15.23	12.45	30.81	10.46	11.58
0.9600000000000001	11.16	11.25	12.61	10.59	9.45	9.83	26.95	31.73	30.77	25.27	29.02	26.77	18.09	18.90	17.46	9.64	10.36	10.50	18.58	17.51	20.13	22.09	19.26	22.68	9.63	10.46	10.48	11.51	11.38	13.83	14.60	13.15	12.60	40.07	37.27	33.57	9.57	8.74	10.11	99.64	125.74	98.55	17.23	15.43	13.57	15.46	nan	nan	13.18	13.22	14.11	11.10	10.84	11.00	17.48	17.37	15.35	9.61	9.41	10.62	18.80	22.23	20.50	19.74	17.44	18.76	28.04	28.98	31.88	18.49	17.59	16.76	6.98	116.67	0.00	30.45	29.77	32.87	37.19	34.78	35.01	6.67	6.67	10.71	1.00	10.00	9.89	30.95	23.72	19.05	2.67	7.29	42.82	25.42	16.49	14.88	25.42	13.30	11.28
0.9700000000000001	12.59	11.40	14.74	12.29	12.59	13.50	29.19	40.89	36.48	28.19	32.27	29.77	23.40	21.27	20.46	12.56	11.74	13.55	23.95	22.57	23.25	22.90	23.79	23.54	13.42	12.52	12.85	13.06	14.11	15.92	14.68	15.14	16.48	43.87	40.36	42.23	11.52	10.41	12.24	128.14	135.09	150.00	20.27	19.74	18.11	16.32	nan	nan	16.14	19.53	16.28	13.37	12.41	12.05	21.51	22.99	24.01	13.19	12.59	10.87	26.52	25.01	27.45	21.71	22.21	22.26	32.77	37.18	35.52	22.36	22.49	24.61	15.22	53.85	44.00	36.58	35.00	38.32	41.18	39.03	42.86	9.38	12.50	9.68	8.91	0.00	10.00	20.61	17.40	26.33	4.92	8.31	41.59	26.86	19.73	18.29	29.41	15.48	15.97
0.9800000000000001	17.34	18.39	17.95	16.41	15.99	17.38	42.63	41.47	47.95	34.23	38.71	41.12	31.61	32.00	30.06	16.71	15.81	16.31	26.97	28.45	29.66	31.50	30.95	32.04	18.02	17.07	18.66	17.32	16.23	18.41	20.27	19.11	21.53	49.46	52.20	55.19	14.18	18.56	15.35	159.20	146.46	183.88	27.11	26.34	26.15	23.02	nan	nan	29.20	26.14	27.24	18.30	18.40	16.44	27.41	31.04	29.57	18.44	18.08	18.37	35.34	36.17	38.87	32.83	35.22	32.68	44.29	44.39	43.18	26.40	31.51	29.92	9.43	50.00	38.89	56.05	48.71	48.66	54.17	54.57	52.27	14.29	11.11	14.71	19.09	36.36	12.55	25.63	25.63	31.93	6.05	12.29	49.53	34.90	24.65	29.22	34.70	22.63	21.01
0.9900000000000001	27.83	25.26	28.72	26.57	31.10	27.18	57.89	75.93	73.99	49.73	58.63	62.36	60.97	60.52	60.36	29.49	28.14	27.96	44.96	44.40	44.20	44.96	46.86	54.64	31.63	31.57	29.27	29.24	26.22	32.46	28.77	35.57	34.86	80.59	88.95	83.68	26.22	24.29	29.35	217.65	206.02	303.10	48.50	42.27	44.88	36.08	nan	nan	50.95	59.27	58.38	32.34	31.91	30.45	49.30	42.69	48.24	30.69	33.41	31.01	63.22	66.93	63.92	60.58	61.42	58.70	77.82	79.94	75.85	57.16	58.74	55.41	17.24	70.00	22.00	85.55	77.30	96.75	91.03	80.54	86.54	20.00	20.00	17.95	52.67	33.33	53.63	31.00	31.00	23.80	0.52	20.80	66.87	55.51	39.38	43.47	49.52	33.15	32.99
1.0	240.51	267.41	391.32	259.47	301.11	533.35	648.04	1203.51	1345.42	594.33	527.51	1133.30	497.57	598.26	1024.15	312.34	309.57	354.09	1513.24	1213.96	971.17	1279.33	1917.74	1118.63	489.03	335.71	553.61	290.76	323.81	313.51	325.17	339.48	631.98	1316.15	1141.04	1229.43	207.27	280.45	399.13	14204.63	13481.40	18086.75	1333.97	1484.83	1776.73	980.90	inf	inf	915.66	780.03	543.95	357.30	334.83	487.90	751.25	866.13	755.80	337.45	326.60	293.87	2431.76	2409.61	1902.47	1076.01	749.39	1415.23	2430.88	1907.56	2723.15	844.75	993.52	956.44	2805.88	4082.35	921.31	1555.13	1877.27	1540.89	6390.92	5822.64	5519.41	539.58	187.50	326.09	301.00	301.00	367.82	512.21	512.21	618.74	18.35	1603.85	1132.18	733.09	771.52	684.69	787.93	1381.89	1284.07

	sachet_3g_6_0	sachet_3g_6_1	sachet_3g_6_2	sachet_3g_6_Others	monthly_2g_7_0	monthly_2g_7_1	monthly_2g_7_2	monthly_2g_7_Others	monthly_2g_8_0	monthly_2g_8_1	monthly_2g_8_Others	sachet_3g_8_0	sachet_3g_8_1	sachet_3g_8_Others	monthly_3g_7_0	monthly_3g_7_1	monthly_3g_7_2	monthly_3g_7_Others	sachet_2g_6_0	sachet_2g_6_1	sachet_2g_6_2	sachet_2g_6_3	sachet_2g_6_4	sachet_2g_6_Others	monthly_2g_6_0	monthly_2g_6_1	monthly_2g_6_2	monthly_2g_6_Others	sachet_2g_7_0	sachet_2g_7_1	sachet_2g_7_2	sachet_2g_7_3	sachet_2g_7_4	sachet_2g_7_5	sachet_2g_7_Others	sachet_3g_7_0	sachet_3g_7_1	sachet_3g_7_2	sachet_3g_7_Others	monthly_3g_8_0	monthly_3g_8_1	monthly_3g_8_2	monthly_3g_8_Others	monthly_3g_6_0	monthly_3g_6_1	monthly_3g_6_2	monthly_3g_6_Others	sachet_2g_8_0	sachet_2g_8_1	sachet_2g_8_2	sachet_2g_8_3	sachet_2g_8_4	sachet_2g_8_5	sachet_2g_8_Others
mobile_number
7000701601	1	0	0	0	1	0	0	0	1	0	0	1	0	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0	1	0	0	0	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	0	0	0
7001524846	1	0	0	0	0	1	0	0	1	0	0	1	0	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0	0	1	0	0	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	0	0	0	1	0	0	0
7002191713	1	0	0	0	1	0	0	0	1	0	0	1	0	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0	1	0	0	0	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	0	0	0	1	0	0	0
7000875565	1	0	0	0	1	0	0	0	1	0	0	1	0	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0	1	0	0	0	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	0	0	0
7000187447	1	0	0	0	1	0	0	0	1	0	0	1	0	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0	1	0	0	0	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	0	0	0

	Churn
0	0.913598
1	0.086402

	onnet_mou_6	onnet_mou_7	onnet_mou_8	offnet_mou_6	offnet_mou_7	offnet_mou_8	loc_og_t2t_mou_6	loc_og_t2t_mou_7	loc_og_t2t_mou_8	loc_og_t2m_mou_6	loc_og_t2m_mou_7	loc_og_t2m_mou_8	loc_og_t2f_mou_6	loc_og_t2f_mou_7	loc_og_t2f_mou_8	loc_og_t2c_mou_8	loc_og_mou_6	loc_og_mou_7	loc_og_mou_8	std_og_t2t_mou_6	std_og_t2t_mou_7	std_og_t2t_mou_8	std_og_t2m_mou_6	std_og_t2m_mou_7	std_og_t2m_mou_8	std_og_t2f_mou_6	std_og_t2f_mou_7	std_og_t2f_mou_8	std_og_mou_6	std_og_mou_7	std_og_mou_8	isd_og_mou_6	isd_og_mou_8	spl_og_mou_8	loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	loc_ic_t2t_mou_8	loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	loc_ic_mou_6	loc_ic_mou_7	loc_ic_mou_8	std_ic_t2t_mou_6	std_ic_t2t_mou_7	std_ic_t2t_mou_8	std_ic_t2m_mou_6	std_ic_t2m_mou_7	std_ic_t2m_mou_8	std_ic_t2f_mou_6	std_ic_t2f_mou_7	std_ic_t2f_mou_8	std_ic_mou_6	std_ic_mou_7	std_ic_mou_8	spl_ic_mou_6	spl_ic_mou_8	isd_ic_mou_6	isd_ic_mou_7	isd_ic_mou_8	ic_others_6	ic_others_7	ic_others_8	total_rech_num_6	total_rech_num_7	total_rech_num_8	max_rech_amt_6	max_rech_amt_7	max_rech_amt_8	last_day_rch_amt_6	last_day_rch_amt_7	last_day_rch_amt_8	aon	Average_rech_amt_6n7	delta_vol_2g	delta_vol_3g	delta_total_og_mou	delta_total_ic_mou	delta_vbc_3g	delta_arpu	delta_total_rech_amt	sachet_3g_6_0	monthly_2g_7_0	monthly_2g_7_1	monthly_2g_7_2	monthly_2g_8_0	monthly_2g_8_1	sachet_3g_8_0	monthly_3g_7_0	monthly_3g_7_1	sachet_2g_6_0	monthly_2g_6_0	monthly_2g_6_1	sachet_2g_7_0	sachet_3g_7_0	monthly_3g_8_0	monthly_3g_8_2	monthly_3g_6_0	monthly_3g_6_1	monthly_3g_6_2	sachet_2g_8_0
0	53.01	52.64	37.48	316.01	195.74	68.36	53.01	52.64	37.48	282.38	171.64	44.51	31.59	17.38	19.43	0.00	366.99	241.68	101.43	0.00	0.00	0.00	0.00	2.11	0.00	2.03	4.59	4.41	2.03	6.71	4.41	0.00	0.00	0.00	18.41	40.79	11.79	292.99	191.98	85.89	6.26	1.21	10.39	317.68	233.99	108.09	0.00	0.00	0.00	0.66	0.00	0.00	5.61	1.53	2.76	6.28	1.53	2.76	0.00	0.00	0.00	0.00	9.55	0.00	0.00	0.00	6.0	5.0	4.0	198.0	198.0	198.0	110.0	130.0	130.0	1423	483.0	-791.7700	1077.750	-202.870	-159.335	71.085	-172.4995	-155.0	1	0	1	0	0	1	1	1	0	1	0	1	1	1	1	0	1	0	0	1
1	91.39	216.14	150.58	504.19	301.98	434.41	40.36	36.21	27.73	37.26	36.73	59.61	0.00	0.00	0.00	0.58	77.63	72.94	87.34	51.03	179.93	122.84	465.96	265.24	356.44	0.00	0.00	0.00	516.99	445.18	479.29	0.96	3.89	14.45	104.39	31.98	35.83	154.11	147.88	243.53	0.00	0.76	0.00	258.51	180.63	279.36	4.03	2.99	0.46	6.36	12.31	3.91	0.00	0.00	0.00	10.39	15.31	4.38	0.58	0.25	19.66	21.96	86.63	0.23	0.56	1.04	8.0	11.0	12.0	110.0	130.0	130.0	0.0	130.0	0.0	189	454.0	0.0000	0.000	28.130	117.745	0.000	48.6160	-94.0	1	1	0	0	1	0	1	1	0	1	1	0	1	1	1	0	1	0	0	1
2	11.96	14.13	0.40	1.51	0.00	0.00	11.96	14.13	0.40	1.51	0.00	0.00	0.00	0.00	0.00	0.00	13.48	14.13	0.40	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	20.58	20.39	97.66	36.84	21.58	18.66	5.48	0.73	1.43	62.91	42.71	117.76	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	5.0	3.0	4.0	252.0	252.0	252.0	252.0	0.0	252.0	2922	403.0	-44.6300	-5.525	-13.405	64.950	0.000	75.3940	151.0	1	1	0	0	1	0	1	0	1	1	1	0	1	1	0	1	0	0	1	1
3	532.66	537.31	738.21	49.03	71.64	39.43	24.46	19.79	37.74	41.26	47.86	39.43	1.19	4.04	0.00	0.00	66.93	71.71	77.18	508.19	517.51	700.46	6.56	18.24	0.00	0.00	1.48	0.00	514.76	537.24	700.46	0.00	0.00	0.00	19.86	28.81	20.24	66.08	94.18	67.54	51.74	68.16	50.08	137.69	191.16	137.88	18.83	14.56	1.28	1.08	20.89	6.83	0.00	3.08	3.05	19.91	38.54	11.16	0.00	0.00	0.00	5.28	7.49	0.00	0.00	0.00	10.0	13.0	12.0	145.0	150.0	145.0	0.0	150.0	0.0	1128	521.0	-10.1500	-108.195	182.315	-39.760	0.000	192.8075	207.0	1	1	0	0	1	0	1	0	1	1	1	0	1	1	0	1	0	1	0	1
4	122.68	105.51	149.33	302.23	211.44	264.11	122.68	105.51	149.33	301.04	194.06	257.14	0.00	0.66	0.51	0.00	423.73	300.24	406.99	0.00	0.00	0.00	1.18	15.75	6.44	0.00	0.96	0.00	1.18	16.71	6.44	0.00	0.00	0.00	228.54	198.24	231.13	412.99	392.98	353.86	81.76	89.69	88.74	723.31	680.93	673.74	0.00	0.00	1.05	8.14	5.33	0.70	11.83	6.58	10.44	19.98	11.91	12.19	0.00	0.00	0.43	0.00	0.48	0.00	0.00	0.00	5.0	5.0	4.0	325.0	154.0	164.0	325.0	154.0	164.0	2453	721.0	654.3125	-686.915	42.505	-31.855	-433.700	-55.1110	-105.0	1	0	0	1	1	0	1	1	0	1	0	1	1	1	0	1	1	0	0	1

	onnet_mou_6	onnet_mou_7	onnet_mou_8	offnet_mou_6	offnet_mou_7	offnet_mou_8	roam_ic_mou_6	roam_ic_mou_7	roam_ic_mou_8	roam_og_mou_6	roam_og_mou_7	roam_og_mou_8	loc_og_t2t_mou_6	loc_og_t2t_mou_7	loc_og_t2t_mou_8	loc_og_t2m_mou_6	loc_og_t2m_mou_7	loc_og_t2m_mou_8	loc_og_t2f_mou_6	loc_og_t2f_mou_7	loc_og_t2f_mou_8	loc_og_t2c_mou_6	loc_og_t2c_mou_7	loc_og_t2c_mou_8	loc_og_mou_6	loc_og_mou_7	loc_og_mou_8	std_og_t2t_mou_6	std_og_t2t_mou_7	std_og_t2t_mou_8	std_og_t2m_mou_6	std_og_t2m_mou_7	std_og_t2m_mou_8	std_og_t2f_mou_6	std_og_t2f_mou_7	std_og_t2f_mou_8	std_og_mou_6	std_og_mou_7	std_og_mou_8	isd_og_mou_6	isd_og_mou_7	isd_og_mou_8	spl_og_mou_6	spl_og_mou_7	spl_og_mou_8	og_others_6	og_others_7	og_others_8	loc_ic_t2t_mou_6	loc_ic_t2t_mou_7	loc_ic_t2t_mou_8	loc_ic_t2m_mou_6	loc_ic_t2m_mou_7	loc_ic_t2m_mou_8	loc_ic_t2f_mou_6	loc_ic_t2f_mou_7	loc_ic_t2f_mou_8	loc_ic_mou_6	loc_ic_mou_7	loc_ic_mou_8	std_ic_t2t_mou_6	std_ic_t2t_mou_7	std_ic_t2t_mou_8	std_ic_t2m_mou_6	std_ic_t2m_mou_7	std_ic_t2m_mou_8	std_ic_t2f_mou_6	std_ic_t2f_mou_7	std_ic_t2f_mou_8	std_ic_mou_6	std_ic_mou_7	std_ic_mou_8	spl_ic_mou_6	spl_ic_mou_7	spl_ic_mou_8	isd_ic_mou_6	isd_ic_mou_7	isd_ic_mou_8	ic_others_6	ic_others_7	ic_others_8	total_rech_num_6	total_rech_num_7	total_rech_num_8	max_rech_amt_6	max_rech_amt_7	max_rech_amt_8	last_day_rch_amt_6	last_day_rch_amt_7	last_day_rch_amt_8	aon	Average_rech_amt_6n7	delta_vol_2g	delta_vol_3g	delta_total_og_mou	delta_total_ic_mou	delta_vbc_3g	delta_arpu	delta_total_rech_amt	sachet_3g_6_0	sachet_3g_6_1	sachet_3g_6_2	monthly_2g_7_0	monthly_2g_7_1	monthly_2g_7_2	monthly_2g_8_0	monthly_2g_8_1	sachet_3g_8_0	sachet_3g_8_1	monthly_3g_7_0	monthly_3g_7_1	monthly_3g_7_2	sachet_2g_6_0	sachet_2g_6_1	sachet_2g_6_2	sachet_2g_6_3	sachet_2g_6_4	monthly_2g_6_0	monthly_2g_6_1	monthly_2g_6_2	sachet_2g_7_0	sachet_2g_7_1	sachet_2g_7_2	sachet_2g_7_3	sachet_2g_7_4	sachet_2g_7_5	sachet_3g_7_0	sachet_3g_7_1	sachet_3g_7_2	monthly_3g_8_0	monthly_3g_8_1	monthly_3g_8_2	monthly_3g_6_0	monthly_3g_6_1	monthly_3g_6_2	sachet_2g_8_0	sachet_2g_8_1	sachet_2g_8_2	sachet_2g_8_3	sachet_2g_8_4	sachet_2g_8_5
count	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.0	38374.0	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00	38374.00
mean	-0.00	-0.00	0.00	0.00	-0.00	-0.00	-0.00	-0.00	0.00	-0.00	-0.00	0.00	-0.00	-0.00	0.00	0.00	-0.00	0.00	0.00	-0.00	-0.00	-0.00	0.00	-0.00	0.00	-0.00	-0.00	-0.00	-0.00	-0.00	-0.00	-0.00	0.00	0.00	0.00	0.00	-0.00	0.00	0.00	-0.00	-0.00	0.00	0.00	0.00	0.00	-0.00	0.0	0.0	0.00	-0.00	-0.00	-0.00	0.00	-0.00	0.00	0.00	0.00	-0.00	-0.00	-0.00	0.00	0.00	-0.00	0.00	0.00	0.00	0.00	0.00	-0.00	-0.00	0.00	-0.00	-0.00	-0.00	-0.00	-0.00	0.00	0.00	-0.00	-0.00	0.00	-0.00	0.00	-0.00	-0.00	-0.00	0.00	0.00	-0.00	0.00	0.00	-0.00	0.00	0.00	-0.00	0.00	-0.00	0.00	-0.00	0.00	-0.00	-0.00	-0.00	-0.00	-0.00	0.00	0.00	-0.00	0.00	0.00	0.00	0.00	0.00	-0.00	0.00	0.00	0.00	0.00	0.00	0.00	-0.00	-0.00	-0.00	0.00	0.00	-0.00	-0.00	0.00	-0.00	-0.00	0.00	-0.00	-0.00	-0.00	-0.00	-0.00	-0.00	0.00	-0.00	0.00	0.00
std	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	0.0	0.0	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00
min	-0.73	-0.68	-0.53	-0.94	-0.89	-0.70	-0.31	-0.32	-0.33	-0.33	-0.36	-0.36	-0.50	-0.49	-0.42	-0.75	-0.73	-0.59	-0.38	-0.38	-0.33	-0.37	-0.37	-0.31	-0.76	-0.74	-0.60	-0.57	-0.54	-0.40	-0.60	-0.57	-0.43	-0.22	-0.22	-0.19	-0.79	-0.74	-0.53	-0.20	-0.18	-0.15	-0.51	-0.53	-0.43	-0.44	0.0	0.0	-0.61	-0.57	-0.48	-0.77	-0.75	-0.61	-0.40	-0.39	-0.35	-0.80	-0.77	-0.63	-0.45	-0.43	-0.33	-0.51	-0.47	-0.39	-0.26	-0.26	-0.22	-0.56	-0.52	-0.41	-0.47	-0.21	-0.20	-0.28	-0.27	-0.22	-0.27	-0.25	-0.22	-1.50	-1.37	-1.02	-1.08	-1.04	-0.86	-0.93	-0.85	-0.64	-1.01	-0.91	-28.14	-27.16	-11.65	-14.66	-22.68	-14.96	-14.68	-3.43	-0.16	-0.08	-3.10	-0.25	-0.09	-3.78	-0.23	-4.62	-0.14	-2.78	-0.23	-0.13	-1.95	-0.22	-0.14	-0.11	-0.09	-2.99	-0.24	-0.08	-1.99	-0.21	-0.14	-0.10	-0.09	-0.08	-3.44	-0.16	-0.07	-3.33	-0.22	-0.12	-2.66	-0.23	-0.12	-2.11	-0.23	-0.14	-0.11	-0.10	-0.09
25%	-0.63	-0.60	-0.52	-0.66	-0.65	-0.66	-0.31	-0.32	-0.33	-0.33	-0.36	-0.36	-0.45	-0.45	-0.42	-0.63	-0.63	-0.59	-0.38	-0.38	-0.33	-0.37	-0.37	-0.31	-0.63	-0.63	-0.60	-0.57	-0.54	-0.40	-0.59	-0.56	-0.43	-0.22	-0.22	-0.19	-0.77	-0.72	-0.53	-0.20	-0.18	-0.15	-0.51	-0.53	-0.43	-0.44	0.0	0.0	-0.53	-0.51	-0.48	-0.62	-0.61	-0.61	-0.40	-0.39	-0.35	-0.63	-0.62	-0.63	-0.45	-0.43	-0.33	-0.49	-0.46	-0.39	-0.26	-0.26	-0.22	-0.51	-0.48	-0.41	-0.47	-0.21	-0.20	-0.28	-0.27	-0.22	-0.27	-0.25	-0.22	-0.68	-0.66	-0.63	-0.40	-0.45	-0.70	-0.64	-0.65	-0.64	-0.73	-0.68	0.12	0.11	-0.41	-0.22	0.08	-0.51	-0.48	0.29	-0.16	-0.08	0.32	-0.25	-0.09	0.26	-0.23	0.22	-0.14	0.36	-0.23	-0.13	0.51	-0.22	-0.14	-0.11	-0.09	0.33	-0.24	-0.08	0.50	-0.21	-0.14	-0.10	-0.09	-0.08	0.29	-0.16	-0.07	0.30	-0.22	-0.12	0.38	-0.23	-0.12	0.47	-0.23	-0.14	-0.11	-0.10	-0.09
50%	-0.42	-0.41	-0.40	-0.33	-0.33	-0.36	-0.31	-0.32	-0.33	-0.33	-0.36	-0.36	-0.32	-0.32	-0.35	-0.37	-0.37	-0.43	-0.38	-0.38	-0.33	-0.37	-0.37	-0.31	-0.36	-0.37	-0.43	-0.50	-0.48	-0.40	-0.45	-0.45	-0.41	-0.22	-0.22	-0.19	-0.42	-0.45	-0.49	-0.20	-0.18	-0.15	-0.43	-0.43	-0.43	-0.44	0.0	0.0	-0.34	-0.33	-0.37	-0.34	-0.35	-0.40	-0.36	-0.35	-0.35	-0.34	-0.35	-0.40	-0.38	-0.36	-0.33	-0.34	-0.34	-0.35	-0.26	-0.26	-0.22	-0.33	-0.33	-0.35	-0.47	-0.21	-0.20	-0.28	-0.27	-0.22	-0.27	-0.25	-0.22	-0.25	-0.30	-0.32	-0.33	-0.25	-0.07	-0.16	-0.34	-0.43	-0.39	-0.35	0.15	0.11	0.23	0.13	0.08	0.08	0.03	0.29	-0.16	-0.08	0.32	-0.25	-0.09	0.26	-0.23	0.22	-0.14	0.36	-0.23	-0.13	0.51	-0.22	-0.14	-0.11	-0.09	0.33	-0.24	-0.08	0.50	-0.21	-0.14	-0.10	-0.09	-0.08	0.29	-0.16	-0.07	0.30	-0.22	-0.12	0.38	-0.23	-0.12	0.47	-0.23	-0.14	-0.11	-0.10	-0.09
75%	0.20	0.15	0.01	0.27	0.26	0.23	-0.27	-0.25	-0.22	-0.28	-0.21	-0.20	0.01	0.00	-0.04	0.23	0.22	0.16	-0.13	-0.13	-0.20	-0.24	-0.21	-0.31	0.24	0.24	0.17	0.08	0.01	-0.20	0.10	0.07	-0.10	-0.22	-0.22	-0.19	0.45	0.39	0.07	-0.20	-0.18	-0.15	0.05	0.07	-0.05	-0.11	0.0	0.0	0.09	0.06	0.03	0.20	0.20	0.17	-0.09	-0.11	-0.15	0.23	0.21	0.20	-0.02	-0.04	-0.14	0.02	-0.00	-0.07	-0.25	-0.26	-0.22	0.05	0.03	-0.04	-0.14	-0.21	-0.20	-0.26	-0.26	-0.22	-0.21	-0.23	-0.22	0.37	0.40	0.28	0.06	0.02	0.22	0.14	0.34	0.58	0.39	0.30	0.15	0.11	0.50	0.36	0.08	0.58	0.58	0.29	-0.16	-0.08	0.32	-0.25	-0.09	0.26	-0.23	0.22	-0.14	0.36	-0.23	-0.13	0.51	-0.22	-0.14	-0.11	-0.09	0.33	-0.24	-0.08	0.50	-0.21	-0.14	-0.10	-0.09	-0.08	0.29	-0.16	-0.07	0.30	-0.22	-0.12	0.38	-0.23	-0.12	0.47	-0.23	-0.14	-0.11	-0.10	-0.09
max	4.09	4.46	5.67	4.02	4.45	5.24	6.11	6.09	6.19	5.41	5.44	5.67	7.06	7.45	7.71	5.26	5.34	5.79	7.25	7.30	7.57	6.32	6.47	7.42	5.35	5.53	5.83	4.02	4.36	5.89	4.04	4.64	5.99	8.74	8.89	9.37	3.56	3.93	5.21	7.82	8.52	10.18	6.06	5.90	6.82	5.40	0.0	0.0	6.65	6.93	7.46	5.33	5.53	5.87	7.24	7.12	7.65	5.24	5.53	5.85	6.78	7.35	8.26	6.96	7.26	7.84	8.18	8.35	9.00	6.70	7.11	7.69	4.48	8.02	7.72	7.97	7.96	9.30	8.11	8.48	9.21	4.09	4.28	4.93	5.68	5.62	6.02	5.45	5.54	5.69	3.66	4.46	4.05	4.24	2.96	3.13	4.48	2.84	2.84	0.29	6.28	12.99	0.32	4.07	11.56	0.26	4.32	0.22	7.05	0.36	4.35	7.92	0.51	4.53	7.29	9.20	11.42	0.33	4.12	12.50	0.50	4.78	7.36	9.71	10.80	11.92	0.29	6.43	13.42	0.30	4.57	8.36	0.38	4.28	8.04	0.47	4.29	7.35	9.30	10.13	10.99

	0.0	0.1	0.2	0.3	0.4	0.5	0.6	0.7	0.8	0.9
mobile_number
7000166926	1	1	1	1	1	0	0	0	0	0
7001343085	1	1	1	0	0	0	0	0	0	0
7001863283	1	1	0	0	0	0	0	0	0	0
7002275981	1	1	1	0	0	0	0	0	0	0
7001086221	1	0	0	0	0	0	0	0	0	0

	sensitivity	specificity	accuracy
0.0	1.000	0.000	0.087
0.1	0.974	0.345	0.399
0.2	0.947	0.523	0.560
0.3	0.910	0.658	0.680
0.4	0.868	0.763	0.772
0.5	0.821	0.834	0.833
0.6	0.770	0.883	0.873
0.7	0.677	0.921	0.899
0.8	0.493	0.953	0.913
0.9	0.234	0.981	0.916

	Column	Rank	Support
99	sachet_3g_6_0	1	True
120	sachet_2g_7_0	1	True
102	monthly_2g_7_0	1	True
135	sachet_2g_8_0	1	True
81	total_rech_num_6	1	True
129	monthly_3g_8_0	1	True
105	monthly_2g_8_0	1	True
83	total_rech_num_8	1	True
117	monthly_2g_6_0	1	True
68	std_ic_t2f_mou_8	1	True
67	std_ic_t2f_mou_7	1	True
112	sachet_2g_6_0	1	True
109	monthly_3g_7_0	1	True
56	loc_ic_t2f_mou_8	1	True
35	std_og_t2f_mou_8	1	True
40	isd_og_mou_7	2	False
53	loc_ic_t2m_mou_8	3	False
19	loc_og_t2f_mou_7	4	False
62	std_ic_t2t_mou_8	5	False
61	std_ic_t2t_mou_7	6	False
107	sachet_3g_8_0	7	False
41	isd_og_mou_8	8	False
89	last_day_rch_amt_8	9	False
11	roam_og_mou_8	10	False
132	monthly_3g_6_0	11	False
39	isd_og_mou_6	12	False
79	ic_others_7	13	False
50	loc_ic_t2t_mou_8	14	False
7	roam_ic_mou_7	15	False
58	loc_ic_mou_7	16	False
71	std_ic_mou_8	17	False
75	isd_ic_mou_6	18	False
33	std_og_t2f_mou_6	19	False
38	std_og_mou_8	20	False
66	std_ic_t2f_mou_6	21	False
29	std_og_t2t_mou_8	22	False
32	std_og_t2m_mou_8	23	False
78	ic_others_6	24	False
44	spl_og_mou_8	25	False
97	delta_arpu	26	False
85	max_rech_amt_7	27	False
70	std_ic_mou_7	28	False
64	std_ic_t2m_mou_7	29	False
30	std_og_t2m_mou_6	30	False
42	spl_og_mou_6	31	False
27	std_og_t2t_mou_6	32	False
18	loc_og_t2f_mou_6	33	False
60	std_ic_t2t_mou_6	34	False
36	std_og_mou_6	35	False
51	loc_ic_t2m_mou_6	36	False
15	loc_og_t2m_mou_6	37	False
94	delta_total_og_mou	38	False
69	std_ic_mou_6	39	False
65	std_ic_t2m_mou_8	40	False
2	onnet_mou_8	41	False
55	loc_ic_t2f_mou_7	42	False
28	std_og_t2t_mou_7	43	False
13	loc_og_t2t_mou_7	44	False
1	onnet_mou_7	45	False
9	roam_og_mou_6	46	False
21	loc_og_t2c_mou_6	47	False
14	loc_og_t2t_mou_8	48	False
84	max_rech_amt_6	49	False
26	loc_og_mou_8	50	False
8	roam_ic_mou_8	51	False
10	roam_og_mou_7	52	False
48	loc_ic_t2t_mou_6	53	False
57	loc_ic_mou_6	54	False
6	roam_ic_mou_6	55	False
106	monthly_2g_8_1	56	False
87	last_day_rch_amt_6	57	False
49	loc_ic_t2t_mou_7	58	False
98	delta_total_rech_amt	59	False
88	last_day_rch_amt_7	60	False
34	std_og_t2f_mou_7	61	False
126	sachet_3g_7_0	62	False
23	loc_og_t2c_mou_8	63	False
103	monthly_2g_7_1	64	False
118	monthly_2g_6_1	65	False
92	delta_vol_2g	66	False
16	loc_og_t2m_mou_7	67	False
4	offnet_mou_7	68	False
43	spl_og_mou_7	69	False
130	monthly_3g_8_1	70	False
20	loc_og_t2f_mou_8	71	False
17	loc_og_t2m_mou_8	72	False
63	std_ic_t2m_mou_6	73	False
93	delta_vol_3g	74	False
76	isd_ic_mou_7	75	False
24	loc_og_mou_6	76	False
12	loc_og_t2t_mou_6	77	False
54	loc_ic_t2f_mou_6	78	False
0	onnet_mou_6	79	False
3	offnet_mou_6	80	False
77	isd_ic_mou_8	81	False
5	offnet_mou_8	82	False
22	loc_og_t2c_mou_7	83	False
95	delta_total_ic_mou	84	False
52	loc_ic_t2m_mou_7	85	False
59	loc_ic_mou_8	86	False
90	aon	87	False
74	spl_ic_mou_8	88	False
136	sachet_2g_8_1	89	False
121	sachet_2g_7_1	90	False
113	sachet_2g_6_1	91	False
108	sachet_3g_8_1	92	False
80	ic_others_8	93	False
137	sachet_2g_8_2	94	False
138	sachet_2g_8_3	95	False
114	sachet_2g_6_2	96	False
123	sachet_2g_7_3	97	False
133	monthly_3g_6_1	98	False
125	sachet_2g_7_5	99	False
131	monthly_3g_8_2	100	False
119	monthly_2g_6_2	101	False
25	loc_og_mou_7	102	False
104	monthly_2g_7_2	103	False
110	monthly_3g_7_1	104	False
100	sachet_3g_6_1	105	False
139	sachet_2g_8_4	106	False
134	monthly_3g_6_2	107	False
111	monthly_3g_7_2	108	False
37	std_og_mou_7	109	False
31	std_og_t2m_mou_7	110	False
140	sachet_2g_8_5	111	False
101	sachet_3g_6_2	112	False
72	spl_ic_mou_6	113	False
86	max_rech_amt_8	114	False
73	spl_ic_mou_7	115	False
96	delta_vbc_3g	116	False
82	total_rech_num_7	117	False
115	sachet_2g_6_3	118	False
124	sachet_2g_7_4	119	False
127	sachet_3g_7_1	120	False
91	Average_rech_amt_6n7	121	False
45	og_others_6	122	False
116	sachet_2g_6_4	123	False
128	sachet_3g_7_2	124	False
122	sachet_2g_7_2	125	False
47	og_others_8	126	False
46	og_others_7	127	False

Dep. Variable:	Churn	No. Observations:	38374
Model:	GLM	Df Residuals:	38358
Model Family:	Binomial	Df Model:	15
Link Function:	logit	Scale:	1.0000
Method:	IRLS	Log-Likelihood:	-19485.
Date:	Mon, 30 Nov 2020	Deviance:	38969.
Time:	21:57:09	Pearson chi2:	2.80e+05
No. Iterations:	7
Covariance Type:	nonrobust

	coef	std err	z	P>\|z\|	[0.025	0.975]
const	-0.2334	0.015	-15.657	0.000	-0.263	-0.204
sachet_3g_6_0	-0.0396	0.014	-2.886	0.004	-0.066	-0.013
sachet_2g_7_0	-0.0980	0.016	-6.201	0.000	-0.129	-0.067
monthly_2g_7_0	0.0096	0.016	0.594	0.552	-0.022	0.041
sachet_2g_8_0	0.0489	0.015	3.359	0.001	0.020	0.077
total_rech_num_6	0.6047	0.017	35.547	0.000	0.571	0.638
monthly_3g_8_0	0.3993	0.017	23.439	0.000	0.366	0.433
monthly_2g_8_0	0.3697	0.018	21.100	0.000	0.335	0.404
total_rech_num_8	-1.2013	0.019	-62.378	0.000	-1.239	-1.164
monthly_2g_6_0	-0.0194	0.015	-1.262	0.207	-0.050	0.011
std_ic_t2f_mou_8	-0.3364	0.026	-12.792	0.000	-0.388	-0.285
std_ic_t2f_mou_7	0.1535	0.019	8.148	0.000	0.117	0.190
sachet_2g_6_0	-0.1117	0.016	-6.847	0.000	-0.144	-0.080
monthly_3g_7_0	-0.2094	0.017	-12.602	0.000	-0.242	-0.177
loc_ic_t2f_mou_8	-1.2743	0.038	-33.599	0.000	-1.349	-1.200
std_og_t2f_mou_8	-0.2476	0.021	-11.621	0.000	-0.289	-0.206

	VIF	P-value
Features
std_ic_t2f_mou_8	1.66	0.0000
sachet_2g_6_0	1.64	0.0000
sachet_2g_7_0	1.57	0.0000
std_ic_t2f_mou_7	1.56	0.0000
monthly_2g_7_0	1.54	0.5524
monthly_3g_7_0	1.54	0.0000
monthly_3g_8_0	1.52	0.0000
monthly_2g_8_0	1.43	0.0000
monthly_2g_6_0	1.38	0.2069
sachet_2g_8_0	1.36	0.0008
total_rech_num_6	1.27	0.0000
total_rech_num_8	1.25	0.0000
std_og_t2f_mou_8	1.20	0.0000
sachet_3g_6_0	1.12	0.0039
loc_ic_t2f_mou_8	1.09	0.0000

	sensitivity	specificity	accuracy
0.0	1.000	0.000	0.500
0.1	0.976	0.224	0.600
0.2	0.947	0.351	0.649
0.3	0.916	0.472	0.694
0.4	0.864	0.598	0.731
0.5	0.794	0.722	0.758
0.6	0.703	0.841	0.772
0.7	0.550	0.930	0.740
0.8	0.310	0.975	0.642
0.9	0.095	0.994	0.544

	PC_1	PC_2	PC_3	PC_4	PC_5	PC_6	PC_7	PC_8	PC_9	PC_10	PC_11	PC_12	PC_13	PC_14	PC_15	PC_16	PC_17	PC_18	PC_19	PC_20	PC_21	PC_22	PC_23	PC_24	PC_25	PC_26	PC_27	PC_28	PC_29	PC_30	PC_31	PC_32	PC_33	PC_34	PC_35	PC_36	PC_37	PC_38	PC_39	PC_40	PC_41	PC_42	PC_43	PC_44	PC_45	Churn
mobile_number
7000166926	-907.572208	-342.923676	13.094442	58.813506	-95.616159	-1050.535219	254.648987	-31.445039	305.140339	-216.814250	95.825021	231.408291	-111.002572	-2.007256	444.977249	31.541681	573.831941	-278.539708	30.768637	-36.915195	-0.293915	-83.574447	-13.960479	-60.930941	-53.208613	56.049658	-17.776675	-12.624526	14.149393	-30.559156	26.064776	-1.080160	-19.814893	-3.293546	-2.717923	7.470255	22.686838	28.696686	-14.312037	4.959030	-8.652543	2.473147	17.080399	-21.824778	-8.062901	0
7001343085	573.898045	-902.385767	-424.839214	-331.153508	-148.987005	-36.955710	-134.445130	265.325388	-92.070929	-164.203586	25.105150	-36.980621	164.785936	-222.908959	-12.573878	-50.569424	-44.767869	-62.984835	-18.100729	-86.239469	-115.399141	-45.776518	16.345395	-21.497140	-10.541281	-71.754047	29.230830	-20.880178	-0.690183	3.220864	-21.223298	65.500636	-39.719437	50.424623	10.586150	43.055219	0.209259	-66.107880	13.583016	25.823444	52.037618	-3.272773	8.493995	19.449057	-38.779466	0
7001863283	-1538.198366	514.032564	846.865497	57.032319	-1126.228705	-84.209511	-44.422495	-88.158881	-58.411887	50.518811	3.052703	-229.100202	-109.215465	-3.253782	7.045279	-85.645393	54.536446	-52.292779	20.978943	-90.806167	96.348659	24.280381	-52.425262	42.430049	-40.627473	-12.715890	-4.331719	-4.092290	50.339358	-0.777645	-35.146663	-121.580965	98.868473	-34.068010	-8.941074	22.920757	1.669933	52.644942	-8.542762	9.087643	-18.403853	3.672076	26.073078	27.246371	19.603368	0
7002275981	486.830772	-224.929803	1130.460535	-496.189015	6.009139	81.106845	-148.667431	170.280911	-7.375197	-99.556793	-159.659135	-14.186219	-98.682096	213.233743	-34.920639	-17.212430	29.644778	4.941994	2.799763	-49.580528	-88.567855	16.809461	-9.471018	4.383889	29.532189	38.211558	32.465761	-5.316497	-60.149577	12.593305	20.988200	80.709846	-50.975160	-3.712583	65.002407	-57.837280	-8.312631	-5.931175	-5.053131	-5.667538	-12.102225	-14.690148	-32.215573	12.517731	-20.158820	0
7001086221	-1420.949314	794.071749	99.221352	155.118564	145.349456	784.723580	-10.947301	609.724272	-172.482377	-42.796400	59.174124	-162.912577	-112.219187	-55.108445	17.303261	-152.111164	-611.929832	181.577435	-211.358075	-77.180329	116.282095	83.488753	-26.254488	128.490023	-69.085253	4.854304	-128.278573	44.328867	-6.470515	-28.782209	14.618174	-31.359379	27.331179	-25.948771	8.941634	-34.840913	-21.933848	17.941556	-0.866531	-19.428832	-5.321193	6.319611	-11.398376	41.907093	-8.296132	0