Table of Contents
The problem of churn is a problem when a number of individuals move out of a collective group. It is one of the main problems that determine the steady-state level of customers in any type of business.
Recently a large number of customers has left Maastricht Telecom. To address this problem Maastricht Telecom provides customer data to solve two important tasks: Descriptive task:
- Characterize loyal and churn customers and propose a focused customer retention program. (This can be done through visualization, descriptive models etc.).
- Predictive task: Find a model that identifies churn customers.
- Select 300 customers using that model from a separate test set and report the number of true churn customers among them.
- Calculate the expected costs for Maastricht Telecom for one month when using your modelon the test set if every customer predicted as churn willget a gift of 10 euro and every true churn customer predicted as loyal will cause a loss of 64 euros (an average month subscription).
To clone and run this application, you'll need Git and Jupyter notebook installed on your computer. From your command line:
# Clone this repository
$ git clone git@github.com:pgupta119/Churn-Problem.git
# Go into the repository
$ cd churn-Problem
# run Churn problem.ipynb file