Data Science Practice: Churn Prediction

Churn Prediction (Synthetic Dataset)

Customer churn, also known as attrition, occurs when a customer stops doing business with a company. Understanding and detecting churn is the first step to retaining these customers and improving the company's offerings. IBM created a dataset for a fictional telecommunication company which can be downloaded here. Conveniently, the dataset contains a column churn to determine whether the customer has churned.

For this problem, you should put yourself in the position of a Data Scientist working at Telco. Your goal is to build a model to predict a customer's likelihood of churn and provide recommendations on how Telco can reduce churn. The solution should include a detailed guide (including code) that makes it possible to replicate the model and recommendations.

In addition to the code and replication guidelines, please also detail:

performance metrics and data used to obtain these evaluations

recommendations for how Telco can use this model to retain customers

findings on the drivers behind churn and how Telco can use these findings to improve their business offerings

reasons behind the algorithm you chose and why you are confident that this is ideal algorithm for this problem

ideas for future improvement

Sample Solution

Note: sample solutions are intended for demonstration purposes (they are not intended to be very detailed). We encourage anyone attempting these problems to dive deeper to answer the questions outlined above.

part 1

part 2