Package used

library(readxl)
Telco <- read_excel("C:/Users/DELL/Desktop/DA/Data.xlsx", 
    sheet = "Main")
View(Telco)
library(ggplot2)
library(ggpie)
library(dplyr)
library(tidyverse)
library(corrplot)
library(lmtest)
library(patchwork)
library(caTools)
library(sjPlot)
library(car)
library(caret)
library(stargazer)
library(rpart)
library(rpart.plot)

Introduction

Data dictionary
Variables Data type
Gender Qualitative data
Age group Qualitative data
Married Qualitative data
Number of Dependents Quantitative data
Number of Referrals Quantitative data
Tenure in Months Time series
Offer Qualitative data
Phone Service Qualitative data
Avg Monthly Long Distance Charges Quantitative data
Multiple Lines Quantitative data
Internet Service Qualitative data
Internet Type Qualitative data
Avg Monthly GB Download Quantitative data
Online Security Qualitative data
Online Backup Qualitative data
Device Protection Plan Qualitative data
Premium Tech Support Qualitative data
Streaming TV Qualitative data
Streaming Movies Qualitative data
Streaming Music Qualitative data
Unlimited Data Qualitative data
Contract Qualitative data
Paperless Billing Qualitative data
Payment Method Qualitative data
Monthly Charge Quantitative data
Total Charges Quantitative data
Total Refunds Quantitative data
Total Extra Data Charges Quantitative data
Total Long Distance Charges Quantitative data
Total Revenue Quantitative data
Customer Status Qualitative data

Overview

## tibble [7,043 × 33] (S3: tbl_df/tbl/data.frame)
##  $ Gender                           : chr [1:7043] "Female" "Male" "Male" "Male" ...
##  $ Cus ID                           : chr [1:7043] "0002-ORFBO" "0003-MKNFE" "0004-TLHLJ" "0011-IGKFF" ...
##  $ Age                              : num [1:7043] 37 46 50 78 75 23 67 52 68 43 ...
##  $ Age Group                        : chr [1:7043] "30 - 59" "30 - 59" "30 - 59" "60 - 89" ...
##  $ Married                          : chr [1:7043] "Yes" "No" "No" "Yes" ...
##  $ Number of Dependents             : num [1:7043] 0 0 0 0 0 3 0 0 0 1 ...
##  $ Number of Referrals              : num [1:7043] 2 0 0 1 3 0 1 8 0 3 ...
##  $ Tenure in Months                 : num [1:7043] 9 9 4 13 3 9 71 63 7 65 ...
##  $ Offer                            : chr [1:7043] "None" "None" "Offer E" "Offer D" ...
##  $ Phone Service                    : chr [1:7043] "Yes" "Yes" "Yes" "Yes" ...
##  $ Avg Monthly Long Distance Charges: num [1:7043] 42.39 10.69 33.65 27.82 7.38 ...
##  $ Multiple Lines                   : chr [1:7043] "No" "Yes" "No" "No" ...
##  $ Internet Service                 : chr [1:7043] "Yes" "Yes" "Yes" "Yes" ...
##  $ Internet Type                    : chr [1:7043] "Cable" "Cable" "Fiber Optic" "Fiber Optic" ...
##  $ Avg Monthly GB Download          : num [1:7043] 16 10 30 4 11 73 14 7 21 14 ...
##  $ Online Security                  : chr [1:7043] "No" "No" "No" "No" ...
##  $ Online Backup                    : chr [1:7043] "Yes" "No" "No" "Yes" ...
##  $ Device Protection Plan           : chr [1:7043] "No" "No" "Yes" "Yes" ...
##  $ Premium Tech Support             : chr [1:7043] "Yes" "No" "No" "No" ...
##  $ Streaming TV                     : chr [1:7043] "Yes" "No" "No" "Yes" ...
##  $ Streaming Movies                 : chr [1:7043] "No" "Yes" "No" "Yes" ...
##  $ Streaming Music                  : chr [1:7043] "No" "Yes" "No" "No" ...
##  $ Unlimited Data                   : chr [1:7043] "Yes" "No" "Yes" "Yes" ...
##  $ Contract                         : chr [1:7043] "One Year" "Month-to-Month" "Month-to-Month" "Month-to-Month" ...
##  $ Paperless Billing                : chr [1:7043] "Yes" "No" "Yes" "Yes" ...
##  $ Payment Method                   : chr [1:7043] "Credit Card" "Credit Card" "Bank Withdrawal" "Bank Withdrawal" ...
##  $ Monthly Charge                   : num [1:7043] 65.6 -4 73.9 98 83.9 ...
##  $ Total Charges                    : num [1:7043] 593 542 281 1238 267 ...
##  $ Total Refunds                    : num [1:7043] 0 38.3 0 0 0 ...
##  $ Total Extra Data Charges         : num [1:7043] 0 10 0 0 0 0 0 20 0 0 ...
##  $ Total Long Distance Charges      : num [1:7043] 381.5 96.2 134.6 361.7 22.1 ...
##  $ Total Revenue                    : num [1:7043] 975 610 415 1600 290 ...
##  $ Customer Status                  : chr [1:7043] "Stayed" "Stayed" "Churned" "Churned" ...
  • Eliminating duplicate values
sum(duplicated(Telco))
## [1] 0

=> No duplicate values

  • Dealing with missing values
sapply(Telco, function(x) sum(is.na(x)))
##                            Gender                            Cus ID 
##                                 0                                 0 
##                               Age                         Age Group 
##                                 0                                 0 
##                           Married              Number of Dependents 
##                                 0                                 0 
##               Number of Referrals                  Tenure in Months 
##                                 0                                 0 
##                             Offer                     Phone Service 
##                                 0                                 0 
## Avg Monthly Long Distance Charges                    Multiple Lines 
##                               682                               682 
##                  Internet Service                     Internet Type 
##                                 0                              1526 
##           Avg Monthly GB Download                   Online Security 
##                              1526                              1526 
##                     Online Backup            Device Protection Plan 
##                              1526                              1526 
##              Premium Tech Support                      Streaming TV 
##                              1526                              1526 
##                  Streaming Movies                   Streaming Music 
##                              1526                              1526 
##                    Unlimited Data                          Contract 
##                              1526                                 0 
##                 Paperless Billing                    Payment Method 
##                                 0                                 0 
##                    Monthly Charge                     Total Charges 
##                                 0                                 0 
##                     Total Refunds          Total Extra Data Charges 
##                                 0                                 0 
##       Total Long Distance Charges                     Total Revenue 
##                                 0                                 0 
##                   Customer Status 
##                                 0
  • Split dataset
telco_use_both = Telco[complete.cases(Telco),] 
#Phone service: Yes/ Internet service: Yes.

miss_inter = c( "Internet Service", 
                "Internet Type", 
                "Avg Monthly GB Download", 
                "Online Security", 
                "Online Backup", 
                "Device Protection Plan", 
                "Premium Tech Support", 
                "Streaming TV", 
                "Streaming Movies", 
                "Streaming Music", 
                "Unlimited Data")
telco_use_phone = Telco[is.na(Telco$`Internet Type`), 
                        -which(names(Telco) %in% miss_inter)]
#Only use phone

miss_phone = c("Phone Service",
               "Avg Monthly Long Distance Charges",
               "Multiple Lines")
telco_use_inter = Telco[is.na(Telco$`Multiple Lines`),
                        -which(names(Telco) %in% miss_phone)]
#Only use internet
  • Extra
quan_val1= c("Number of Dependents",
              "Number of Referrals",
              "Avg Monthly Long Distance Charges",
              "Avg Monthly GB Download",
              "Monthly Charge",
              "Total Revenue")
quan_val2= c("Number of Dependents",
              "Number of Referrals",
              "Avg Monthly Long Distance Charges",
              "Monthly Charge",
              "Total Revenue")
quan_val3= c("Number of Dependents",
              "Number of Referrals",
              "Avg Monthly GB Download",
              "Monthly Charge",
              "Total Revenue")

end_of_quarter = my("06-2022")
# Calculate join date
telco_use_both$Join_Date = end_of_quarter - months(telco_use_both$`Tenure in Months`)
telco_use_both$Join_Date = format(telco_use_both$Join_Date, "%Y-%m")

#Churn
telco_use_both$Churn = case_when( 
  telco_use_both$`Customer Status`== "Churned" ~ "Yes", 
  TRUE ~ "No")

Descriptive statistics

Univariate analysis

  • Summary statistics
summary(telco_use_both)
##     Gender             Cus ID               Age         Age Group        
##  Length:4835        Length:4835        Min.   :19.00   Length:4835       
##  Class :character   Class :character   1st Qu.:33.00   Class :character  
##  Mode  :character   Mode  :character   Median :47.00   Mode  :character  
##                                        Mean   :47.76                     
##                                        3rd Qu.:62.00                     
##                                        Max.   :80.00                     
##    Married          Number of Dependents Number of Referrals Tenure in Months
##  Length:4835        Min.   :0.000        Min.   : 0.000      Min.   : 1.00   
##  Class :character   1st Qu.:0.000        1st Qu.: 0.000      1st Qu.: 9.00   
##  Mode  :character   Median :0.000        Median : 0.000      Median :30.00   
##                     Mean   :0.376        Mean   : 1.895      Mean   :33.04   
##                     3rd Qu.:0.000        3rd Qu.: 3.000      3rd Qu.:56.00   
##                     Max.   :8.000        Max.   :11.000      Max.   :72.00   
##     Offer           Phone Service      Avg Monthly Long Distance Charges
##  Length:4835        Length:4835        Min.   : 1.01                    
##  Class :character   Class :character   1st Qu.:13.04                    
##  Mode  :character   Mode  :character   Median :25.82                    
##                                        Mean   :25.56                    
##                                        3rd Qu.:37.97                    
##                                        Max.   :49.99                    
##  Multiple Lines     Internet Service   Internet Type     
##  Length:4835        Length:4835        Length:4835       
##  Class :character   Class :character   Class :character  
##  Mode  :character   Mode  :character   Mode  :character  
##                                                          
##                                                          
##                                                          
##  Avg Monthly GB Download Online Security    Online Backup     
##  Min.   : 2.00           Length:4835        Length:4835       
##  1st Qu.:13.00           Class :character   Class :character  
##  Median :21.00           Mode  :character   Mode  :character  
##  Mean   :26.09                                                
##  3rd Qu.:30.00                                                
##  Max.   :85.00                                                
##  Device Protection Plan Premium Tech Support Streaming TV      
##  Length:4835            Length:4835          Length:4835       
##  Class :character       Class :character     Class :character  
##  Mode  :character       Mode  :character     Mode  :character  
##                                                                
##                                                                
##                                                                
##  Streaming Movies   Streaming Music    Unlimited Data       Contract        
##  Length:4835        Length:4835        Length:4835        Length:4835       
##  Class :character   Class :character   Class :character   Class :character  
##  Mode  :character   Mode  :character   Mode  :character   Mode  :character  
##                                                                             
##                                                                             
##                                                                             
##  Paperless Billing  Payment Method     Monthly Charge   Total Charges   
##  Length:4835        Length:4835        Min.   :-10.00   Min.   :  42.9  
##  Class :character   Class :character   1st Qu.: 69.40   1st Qu.: 659.5  
##  Mode  :character   Mode  :character   Median : 81.70   Median :2347.9  
##                                        Mean   : 80.31   Mean   :2901.1  
##                                        3rd Qu.: 95.60   3rd Qu.:4870.3  
##                                        Max.   :118.75   Max.   :8684.8  
##  Total Refunds    Total Extra Data Charges Total Long Distance Charges
##  Min.   : 0.000   Min.   :  0.000          Min.   :   1.13            
##  1st Qu.: 0.000   1st Qu.:  0.000          1st Qu.: 139.97            
##  Median : 0.000   Median :  0.000          Median : 517.80            
##  Mean   : 2.081   Mean   :  8.695          Mean   : 848.09            
##  3rd Qu.: 0.000   3rd Qu.:  0.000          3rd Qu.:1360.96            
##  Max.   :49.570   Max.   :150.000          Max.   :3536.64            
##  Total Revenue      Customer Status     Join_Date            Churn          
##  Min.   :   46.92   Length:4835        Length:4835        Length:4835       
##  1st Qu.:  899.48   Class :character   Class :character   Class :character  
##  Median : 3081.23   Mode  :character   Mode  :character   Mode  :character  
##  Mean   : 3755.82                                                           
##  3rd Qu.: 6284.90                                                           
##  Max.   :11979.34
sapply(telco_use_both, function(x) sd(x))
##                            Gender                            Cus ID 
##                                NA                                NA 
##                               Age                         Age Group 
##                        17.3026737                                NA 
##                           Married              Number of Dependents 
##                                NA                         0.8756364 
##               Number of Referrals                  Tenure in Months 
##                         2.9407402                        24.6352469 
##                             Offer                     Phone Service 
##                                NA                                NA 
## Avg Monthly Long Distance Charges                    Multiple Lines 
##                        14.2414400                                NA 
##                  Internet Service                     Internet Type 
##                                NA                                NA 
##           Avg Monthly GB Download                   Online Security 
##                        19.5609900                                NA 
##                     Online Backup            Device Protection Plan 
##                                NA                                NA 
##              Premium Tech Support                      Streaming TV 
##                                NA                                NA 
##                  Streaming Movies                   Streaming Music 
##                                NA                                NA 
##                    Unlimited Data                          Contract 
##                                NA                                NA 
##                 Paperless Billing                    Payment Method 
##                                NA                                NA 
##                    Monthly Charge                     Total Charges 
##                        21.3796392                      2415.4266589 
##                     Total Refunds          Total Extra Data Charges 
##                         8.1352920                        28.0598134 
##       Total Long Distance Charges                     Total Revenue 
##                       864.4759327                      3087.3517910 
##                   Customer Status                         Join_Date 
##                                NA                                NA 
##                             Churn 
##                                NA
  • Demographic

  • Customer segments

  • Quantitative data

Multivariate analysis

  • Customer segments
by(telco_use_both$`Total Revenue`,telco_use_both$Offer,summary)
## telco_use_both$Offer: None
##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##    52.92   861.99  2960.85  3622.36  5958.91 11688.90 
## ------------------------------------------------------------ 
## telco_use_both$Offer: Offer A
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    4229    7188    8346    8321    9465   11979 
## ------------------------------------------------------------ 
## telco_use_both$Offer: Offer B
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    2514    5102    6072    6143    7148   10185 
## ------------------------------------------------------------ 
## telco_use_both$Offer: Offer C
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    1308    2700    3245    3361    3971    6028 
## ------------------------------------------------------------ 
## telco_use_both$Offer: Offer D
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   419.3  1209.4  1561.2  1633.8  1980.6  3273.1 
## ------------------------------------------------------------ 
## telco_use_both$Offer: Offer E
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   46.92  114.80  273.58  371.48  556.04 1351.16

  • Revenue

  • Churn analysis

Correlation

Modeling

Function

  • Logistic model

\[ p=P(y=1)=\frac{\epsilon^{\beta_1 + \beta_2*X_2 + ... + \beta_n*X_n}}{1+\epsilon^{\beta_1 + \beta_2*X_2 + ... + \beta_n*X_n}} \]

  • Logit form

\[ ln(\frac{p}{1-p}) = \beta_1 + \beta_2*X_2 + \beta_3*X_3 + ... + \beta_n*X_n \]

When \(X_j\) changes 1 unit => odd ratio \(\frac{p}{1-p}\) changes \(\epsilon^{\beta_j}\) times

Logistic regression

  • Split data
telco_use_both = mutate(telco_use_both, 
                        Churn = ifelse(Churn == "Yes", 1, 0))
telco_use_both$Churn_text = ifelse(telco_use_both$Churn == 0, 
                                   "Stayed","Churned")
set.seed(2610)
split = sample.split(telco_use_both$Churn,SplitRatio = 0.8)
train = subset(telco_use_both, split == TRUE)
test  = subset(telco_use_both, split == FALSE)

train1 = subset(train, 
                select = - c(`Cus ID`, Age, `Phone Service`,
                      `Internet Service`, `Join_Date`,`Customer Status`,
                      `Total Charges`, `Total Refunds`,
                      `Total Extra Data Charges`,
                      `Total Long Distance Charges`,Churn_text))
train2 = subset(train, 
                select = - c(`Cus ID`, Age, `Phone Service`,
                      `Internet Service`, `Join_Date`,`Customer Status`,
                      `Total Charges`, `Total Refunds`,
                      `Total Extra Data Charges`,
                      `Total Long Distance Charges`,Churn))
test1 = subset(test, 
                select = - c(`Cus ID`, Age, `Phone Service`,
                      `Internet Service`, `Join_Date`,`Customer Status`,
                      `Total Charges`, `Total Refunds`,
                      `Total Extra Data Charges`,
                      `Total Long Distance Charges`,Churn_text))
test2 = subset(test, 
                select = - c(`Cus ID`, Age, `Phone Service`,
                      `Internet Service`, `Join_Date`,`Customer Status`,
                      `Total Charges`, `Total Refunds`,
                      `Total Extra Data Charges`,
                      `Total Long Distance Charges`,Churn))
  • Logistic regression model
md1 = glm(Churn ~ ., data = train1, family = "binomial")

md2 = glm(Churn ~ . + Gender*`Total Revenue` +
            `Number of Dependents`*`Unlimited Data`+
            `Contract`*`Avg Monthly GB Download`+
            I(`Tenure in Months`^2),
          data = train1, family = "binomial")

md3 = glm(Churn ~ . + Gender*`Total Revenue` +
            `Number of Dependents`*`Unlimited Data`+
            `Contract`*`Avg Monthly GB Download`+
            I(`Tenure in Months`^2) - Offer,
          data = train1, family = "binomial")
## 
## ===========================================================================
##                                                  Dependent variable:       
##                                            --------------------------------
##                                                         Churn              
##                                               (1)        (2)        (3)    
## ---------------------------------------------------------------------------
## GenderMale                                   -0.091    -0.285**   -0.277** 
##                                             (0.091)    (0.129)    (0.129)  
##                                                                            
## `Age Group`30 - 59                           0.323*     0.291      0.273   
##                                             (0.191)    (0.190)    (0.189)  
##                                                                            
## `Age Group`60 - 89                          0.911***   0.906***   0.896*** 
##                                             (0.214)    (0.213)    (0.212)  
##                                                                            
## MarriedYes                                  1.648***   1.656***   1.648*** 
##                                             (0.139)    (0.139)    (0.139)  
##                                                                            
## `Number of Dependents`                     -0.787***   -0.326*    -0.317*  
##                                             (0.088)    (0.172)    (0.171)  
##                                                                            
## `Number of Referrals`                      -0.522***  -0.520***  -0.517*** 
##                                             (0.041)    (0.041)    (0.040)  
##                                                                            
## `Tenure in Months`                         -0.072***  -0.090***  -0.101*** 
##                                             (0.012)    (0.013)    (0.012)  
##                                                                            
## OfferOffer A                                0.909***    0.455              
##                                             (0.302)    (0.322)             
##                                                                            
## OfferOffer B                                 -0.302     -0.304             
##                                             (0.191)    (0.190)             
##                                                                            
## OfferOffer C                                -0.346*     -0.135             
##                                             (0.196)    (0.208)             
##                                                                            
## OfferOffer D                               -0.566***  -0.468***            
##                                             (0.155)    (0.159)             
##                                                                            
## OfferOffer E                                 0.113      -0.002             
##                                             (0.141)    (0.148)             
##                                                                            
## `Avg Monthly Long Distance Charges`        -0.011***   -0.008**   -0.008** 
##                                             (0.004)    (0.004)    (0.004)  
##                                                                            
## `Multiple Lines`Yes                         0.234**    0.267**    0.257**  
##                                             (0.106)    (0.108)    (0.107)  
##                                                                            
## `Internet Type`DSL                          -0.366**   -0.331*    -0.334*  
##                                             (0.172)    (0.173)    (0.173)  
##                                                                            
## `Internet Type`Fiber Optic                   -0.084     -0.002     0.004   
##                                             (0.180)    (0.183)    (0.183)  
##                                                                            
## `Avg Monthly GB Download`                    0.004      0.001      0.001   
##                                             (0.004)    (0.004)    (0.004)  
##                                                                            
## `Online Security`Yes                       -0.417***  -0.419***  -0.437*** 
##                                             (0.113)    (0.114)    (0.114)  
##                                                                            
## `Online Backup`Yes                         -0.301***  -0.297***  -0.290*** 
##                                             (0.107)    (0.108)    (0.108)  
##                                                                            
## `Device Protection Plan`Yes                  -0.060     -0.021     -0.018  
##                                             (0.107)    (0.109)    (0.108)  
##                                                                            
## `Premium Tech Support`Yes                  -0.426***  -0.410***  -0.398*** 
##                                             (0.115)    (0.116)    (0.115)  
##                                                                            
## `Streaming TV`Yes                            0.144      0.173      0.175   
##                                             (0.118)    (0.119)    (0.119)  
##                                                                            
## `Streaming Movies`Yes                        -0.227     -0.181     -0.154  
##                                             (0.181)    (0.184)    (0.183)  
##                                                                            
## `Streaming Music`Yes                        0.475***   0.477***   0.455*** 
##                                             (0.172)    (0.174)    (0.173)  
##                                                                            
## `Unlimited Data`Yes                          0.013      0.093      0.090   
##                                             (0.127)    (0.132)    (0.132)  
##                                                                            
## ContractOne Year                           -1.204***  -1.612***  -1.607*** 
##                                             (0.146)    (0.218)    (0.217)  
##                                                                            
## ContractTwo Year                           -2.914***  -3.407***  -3.343*** 
##                                             (0.245)    (0.363)    (0.360)  
##                                                                            
## `Paperless Billing`Yes                      0.256**    0.264**    0.259**  
##                                             (0.105)    (0.106)    (0.105)  
##                                                                            
## `Payment Method`Credit Card                -0.339***  -0.334***  -0.327*** 
##                                             (0.106)    (0.107)    (0.106)  
##                                                                            
## `Payment Method`Mailed Check                0.595***   0.619***   0.596*** 
##                                             (0.219)    (0.222)    (0.219)  
##                                                                            
## `Monthly Charge`                            0.012***   0.012***   0.012*** 
##                                             (0.004)    (0.004)    (0.004)  
##                                                                            
## `Total Revenue`                            0.0004***   0.0002**   0.0002** 
##                                             (0.0001)   (0.0001)   (0.0001) 
##                                                                            
## I(`Tenure in Months`2)                                 0.001***   0.001*** 
##                                                        (0.0002)   (0.0001) 
##                                                                            
## GenderMale:`Total Revenue`                             0.0001**   0.0001** 
##                                                       (0.00003)  (0.00003) 
##                                                                            
## `Number of Dependents`:`Unlimited Data`Yes            -0.566***  -0.579*** 
##                                                        (0.195)    (0.194)  
##                                                                            
## `Avg Monthly GB Download`:ContractOne Year             0.015**    0.015**  
##                                                        (0.006)    (0.006)  
##                                                                            
## `Avg Monthly GB Download`:ContractTwo Year              0.015      0.015   
##                                                        (0.011)    (0.011)  
##                                                                            
## Constant                                     -0.350     -0.156     -0.118  
##                                             (0.404)    (0.411)    (0.406)  
##                                                                            
## ---------------------------------------------------------------------------
## Observations                                 3,868      3,868      3,868   
## Log Likelihood                             -1,500.350 -1,485.281 -1,491.863
## Akaike Inf. Crit.                          3,066.699  3,046.562  3,049.727 
## ===========================================================================
## Note:                                           *p<0.1; **p<0.05; ***p<0.01
  • Best model
anova(md1,md2,test = "LR")
## Analysis of Deviance Table
## 
## Model 1: Churn ~ Gender + `Age Group` + Married + `Number of Dependents` + 
##     `Number of Referrals` + `Tenure in Months` + Offer + `Avg Monthly Long Distance Charges` + 
##     `Multiple Lines` + `Internet Type` + `Avg Monthly GB Download` + 
##     `Online Security` + `Online Backup` + `Device Protection Plan` + 
##     `Premium Tech Support` + `Streaming TV` + `Streaming Movies` + 
##     `Streaming Music` + `Unlimited Data` + Contract + `Paperless Billing` + 
##     `Payment Method` + `Monthly Charge` + `Total Revenue`
## Model 2: Churn ~ Gender + `Age Group` + Married + `Number of Dependents` + 
##     `Number of Referrals` + `Tenure in Months` + Offer + `Avg Monthly Long Distance Charges` + 
##     `Multiple Lines` + `Internet Type` + `Avg Monthly GB Download` + 
##     `Online Security` + `Online Backup` + `Device Protection Plan` + 
##     `Premium Tech Support` + `Streaming TV` + `Streaming Movies` + 
##     `Streaming Music` + `Unlimited Data` + Contract + `Paperless Billing` + 
##     `Payment Method` + `Monthly Charge` + `Total Revenue` + Gender * 
##     `Total Revenue` + `Number of Dependents` * `Unlimited Data` + 
##     Contract * `Avg Monthly GB Download` + I(`Tenure in Months`^2)
##   Resid. Df Resid. Dev Df Deviance  Pr(>Chi)    
## 1      3835     3000.7                          
## 2      3830     2970.6  5   30.137 1.386e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
anova(md2,md3,test = "LR")
## Analysis of Deviance Table
## 
## Model 1: Churn ~ Gender + `Age Group` + Married + `Number of Dependents` + 
##     `Number of Referrals` + `Tenure in Months` + Offer + `Avg Monthly Long Distance Charges` + 
##     `Multiple Lines` + `Internet Type` + `Avg Monthly GB Download` + 
##     `Online Security` + `Online Backup` + `Device Protection Plan` + 
##     `Premium Tech Support` + `Streaming TV` + `Streaming Movies` + 
##     `Streaming Music` + `Unlimited Data` + Contract + `Paperless Billing` + 
##     `Payment Method` + `Monthly Charge` + `Total Revenue` + Gender * 
##     `Total Revenue` + `Number of Dependents` * `Unlimited Data` + 
##     Contract * `Avg Monthly GB Download` + I(`Tenure in Months`^2)
## Model 2: Churn ~ Gender + `Age Group` + Married + `Number of Dependents` + 
##     `Number of Referrals` + `Tenure in Months` + Offer + `Avg Monthly Long Distance Charges` + 
##     `Multiple Lines` + `Internet Type` + `Avg Monthly GB Download` + 
##     `Online Security` + `Online Backup` + `Device Protection Plan` + 
##     `Premium Tech Support` + `Streaming TV` + `Streaming Movies` + 
##     `Streaming Music` + `Unlimited Data` + Contract + `Paperless Billing` + 
##     `Payment Method` + `Monthly Charge` + `Total Revenue` + Gender * 
##     `Total Revenue` + `Number of Dependents` * `Unlimited Data` + 
##     Contract * `Avg Monthly GB Download` + I(`Tenure in Months`^2) - 
##     Offer
##   Resid. Df Resid. Dev Df Deviance Pr(>Chi)  
## 1      3830     2970.6                       
## 2      3835     2983.7 -5  -13.164  0.02189 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

=> Choose model 3

  • Coefficients results
exp(coef(md3))
##                                (Intercept) 
##                                 0.88885446 
##                                 GenderMale 
##                                 0.75768636 
##                         `Age Group`30 - 59 
##                                 1.31383579 
##                         `Age Group`60 - 89 
##                                 2.44987857 
##                                 MarriedYes 
##                                 5.19869541 
##                     `Number of Dependents` 
##                                 0.72827531 
##                      `Number of Referrals` 
##                                 0.59622598 
##                         `Tenure in Months` 
##                                 0.90427703 
##        `Avg Monthly Long Distance Charges` 
##                                 0.99187532 
##                        `Multiple Lines`Yes 
##                                 1.29266186 
##                         `Internet Type`DSL 
##                                 0.71621831 
##                 `Internet Type`Fiber Optic 
##                                 1.00405994 
##                  `Avg Monthly GB Download` 
##                                 1.00069004 
##                       `Online Security`Yes 
##                                 0.64603932 
##                         `Online Backup`Yes 
##                                 0.74804590 
##                `Device Protection Plan`Yes 
##                                 0.98193415 
##                  `Premium Tech Support`Yes 
##                                 0.67171692 
##                          `Streaming TV`Yes 
##                                 1.19144175 
##                      `Streaming Movies`Yes 
##                                 0.85713362 
##                       `Streaming Music`Yes 
##                                 1.57653697 
##                        `Unlimited Data`Yes 
##                                 1.09413278 
##                           ContractOne Year 
##                                 0.20048895 
##                           ContractTwo Year 
##                                 0.03534785 
##                     `Paperless Billing`Yes 
##                                 1.29569950 
##                `Payment Method`Credit Card 
##                                 0.72129729 
##               `Payment Method`Mailed Check 
##                                 1.81477815 
##                           `Monthly Charge` 
##                                 1.01217938 
##                            `Total Revenue` 
##                                 1.00021258 
##                    I(`Tenure in Months`^2) 
##                                 1.00069977 
##                 GenderMale:`Total Revenue` 
##                                 1.00006928 
## `Number of Dependents`:`Unlimited Data`Yes 
##                                 0.56019747 
## `Avg Monthly GB Download`:ContractOne Year 
##                                 1.01501412 
## `Avg Monthly GB Download`:ContractTwo Year 
##                                 1.01513211
  • Plot coefficients

  • Prediction
pr1 = predict(md3, test1, type = "response")
predicted = round(pr1)
tbl1 = table(Prediction = predicted, Actual = test1$Churn)
tbl1
##           Actual
## Prediction   0   1
##          0 547  96
##          1 103 221
spec = 547/(547+103)*100 #Specificity
sens = 221/(221+96)*100  #Sensitivity
FP   = 100 - sens        #False Positive
FN   = 100 - spec        #False Negative
spec #Specificity      84.15%
## [1] 84.15385
sens #Sensitivity      69.72%
## [1] 69.71609
FP   #False Positive   30.28%
## [1] 30.28391
FN   #False Negative   15.85%
## [1] 15.84615

84.15% of customers are predicted “Stayed” and the result is true negative which means predicted correct.

69.72% of customers are predicted “Churned” and the result is true positive which means predicted correct.

30.28% of customers have churned but the result is predicted “Stayed”

15.85% of customers have stayed but the result is predicted “Churned”

#Note: ?confusionMatrix
confusionMatrix(tbl1, positive = "1", mode = "everything")
## Confusion Matrix and Statistics
## 
##           Actual
## Prediction   0   1
##          0 547  96
##          1 103 221
##                                           
##                Accuracy : 0.7942          
##                  95% CI : (0.7673, 0.8193)
##     No Information Rate : 0.6722          
##     P-Value [Acc > NIR] : <2e-16          
##                                           
##                   Kappa : 0.5357          
##                                           
##  Mcnemar's Test P-Value : 0.6706          
##                                           
##             Sensitivity : 0.6972          
##             Specificity : 0.8415          
##          Pos Pred Value : 0.6821          
##          Neg Pred Value : 0.8507          
##               Precision : 0.6821          
##                  Recall : 0.6972          
##                      F1 : 0.6895          
##              Prevalence : 0.3278          
##          Detection Rate : 0.2285          
##    Detection Prevalence : 0.3351          
##       Balanced Accuracy : 0.7693          
##                                           
##        'Positive' Class : 1               
## 

Decision tree

dtr = rpart(Churn_text ~ ., data = train2, method = "class")

Initially, probability of stay decision is 0.67, first sample size divides into 57% and 43%. In layer 2, probability of stay decision is 0.49, 57% out of first sample size continues to divide into 48% and 9%.

pr2 = predict(dtr, test2, type = "class")
tbl2= table(Prediction = pr2, Actual = test2$Churn_text)
#Note: ?confusionMatrix
confusionMatrix(tbl2, positive = "Churned", mode = "everything")
## Confusion Matrix and Statistics
## 
##           Actual
## Prediction Churned Stayed
##    Churned     200     92
##    Stayed      117    558
##                                           
##                Accuracy : 0.7839          
##                  95% CI : (0.7566, 0.8094)
##     No Information Rate : 0.6722          
##     P-Value [Acc > NIR] : 1.149e-14       
##                                           
##                   Kappa : 0.4995          
##                                           
##  Mcnemar's Test P-Value : 0.09689         
##                                           
##             Sensitivity : 0.6309          
##             Specificity : 0.8585          
##          Pos Pred Value : 0.6849          
##          Neg Pred Value : 0.8267          
##               Precision : 0.6849          
##                  Recall : 0.6309          
##                      F1 : 0.6568          
##              Prevalence : 0.3278          
##          Detection Rate : 0.2068          
##    Detection Prevalence : 0.3020          
##       Balanced Accuracy : 0.7447          
##                                           
##        'Positive' Class : Churned         
## 

Diagnostic test

vif(md3, type ="predictor")
##                                              GVIF Df GVIF^(1/(2*Df))
## Gender                                   2.012650  1        1.418679
## `Age Group`                              3.227297  2        1.340324
## Married                                  2.066820  1        1.437644
## `Number of Dependents`                   4.347828  1        2.085145
## `Number of Referrals`                    1.762788  1        1.327700
## `Tenure in Months`                      32.719743  1        5.720117
## `Avg Monthly Long Distance Charges`      1.669636  1        1.292144
## `Multiple Lines`                         1.392603  1        1.180086
## `Internet Type`                          2.494508  2        1.256742
## `Avg Monthly GB Download`                2.999940  1        1.732034
## `Online Security`                        1.181064  1        1.086768
## `Online Backup`                          1.282129  1        1.132311
## `Device Protection Plan`                 1.294678  1        1.137839
## `Premium Tech Support`                   1.205969  1        1.098166
## `Streaming TV`                           1.702718  1        1.304882
## `Streaming Movies`                       4.055482  1        2.013823
## `Streaming Music`                        3.552718  1        1.884866
## `Unlimited Data`                         1.115683  1        1.056259
## Contract                                 8.871002  2        1.725811
## `Paperless Billing`                      1.070768  1        1.034779
## `Payment Method`                         1.084166  2        1.020408
## `Monthly Charge`                         4.081645  1        2.020308
## `Total Revenue`                         43.994619  1        6.632844
## I(`Tenure in Months`^2)                 17.120265  1        4.137664
## Gender:`Total Revenue`                   3.153862  1        1.775912
## `Number of Dependents`:`Unlimited Data`  4.297086  1        2.072941
## `Avg Monthly GB Download`:Contract       8.313728  2        1.698044