You are on page 1of 23

CUSTOMER CHURN

ANALYSIS & PREDICTION


CONTENT
 Introduction
 Problem Statement
 Project objective
 Dataset description
 Churn prediction model
 Features present in churn prediction
 Data Exploration
 Models implementation
I. Churn prediction using logistic regression
II. Churn prediction using multiple logistic regression
III. Churn prediction using decision tree
IV. Churn prediction using random forest classifier.
 Churn prediction models comparison
INTRODUCTION
• Customer churn refers to that customer switch over from
one service provider to another service provider.

• Churn is a challenging problem for a subscription service


provider.

• The main goal of those companies to retain their existing


customers.

• Churn prediction model can be used to analyze the


historical data available within the business organization
to identify the customers who are at high risk of churning
PROBLEM STATEMENT
Customer churn refers to when a customer switches from
one service provider to another. Churn is a problem for
any provider of a subscription service or recurring
purchasable.
This is where the churn prediction model can help the
business to identify such high risk customers and thereby
helps in maintaining the existing base and increase in
revenues.
Churn prediction is also important because of the fact that
acquiring new customers is much costly than retaining the
existing one.
PROJECT OBJECTIVE
 To predict Customer Churn.

Highlighting the main variables/factors influencing


Customer Churn.

Use various ML algorithms to build prediction models,


evaluating the accuracy and performance of these
models.

Finding out the best model for our business case &
proving executive summary
DATASET DESCRIPTION
• Source dataset is in csv format.

• Dataset contains 7043 rows and 14 columns.

• There is no missing values for the provided input dataset.

• Churn is the variable which notifies whether a particular


customer is churned or not.
CHURN PREDICTION MODEL
FEATURES PRESENT IN CHURN PREDICTION
• Senior Citizen int64
• Monthly Charges float64
• Total charges object
• Gender object
• Partner object
• Dependents object
• Phone Service object
• Multiple Lines object
• Internet Service object
• OnlineSecutiry object
• Device Protection object
• Online Backup object
• TechSupport object
• Streaming TV Streaming Movies object
• Contract object
• Paperless Billing object
• Payment Method object
• Tenure int64
DATA EXPLORATION
• UNIVARIATE ANALYSIS
Relationship between Monthly Charges and Total Charges
• Build a correlation of all predictors with 'Churn'
• BIVARIATE ANALYSIS
MODEL IMPLEMENT
The dataset is trained and tested with 3 methods

I. Churn prediction using logistic regression


II. Churn prediction using multiple logistic regression
III. Churn prediction using decision tree
IV. Churn prediction using random forest classifier
CHURN PREDICTION USING LOGISTIC
REGRESSION
Build a simple logistic regression model where dependent
variable is ‘Churn’ & independent variable is “Monthly
Charges”.

 Divide the dataset in 65:35 ratio.

 Build the model on train set and predict the values on test
set.

 Build the confusion matrix and get the accuracy score.


CHURN PREDICTION USING MULTIPLE
LOGISTIC REGRESSION
Build a multiple logistic regression model where dependent
variable is ‘Churn’ & independent variable are “tenure”
“Monthly Charges”.

 Divide the dataset in 80:20 ratio.

 Build the model on train set and predict the values on test
set.

 Build the confusion matrix and get the accuracy score


CHURN PREDICTION USING DECISION TREE
Build a decision tree model where dependent variable is
‘Churn’ & independent variable are “tenure”.

 Divide the dataset in 80:20 ratio.

 Build the model on train set and predict the values on test
set.

 Build the confusion matrix and calculate the accuracy.


CHURN PREDICTION USING RANDOM FOREST
CLASSIFIER
Build a Random forest model where dependent variable is
‘Churn’ & independent variable are “tenure” and “Monthly
Charges”

 Divide the dataset in 70:30 ratio.

 Build the model on train set and predict the values on test
set.

 Build the confusion matrix and calculate the accuracy


CHURN PREDICTION MODEL COMPARISION

We conclude that from the above accuracy table, that Multiple logistic regression is best
model for the Churn prediction problem.
THANK YOU

You might also like