For a business with various services, in this case, a fictional Telco company, it is essential to keep old customers because studies suggest that depending on the industry you are in, acquiring a new customer can cost five to seven times more than retaining an old one. Therefore, I created this machine learning model to predict customer churn, which promises to help businesses make better customer retention programs.

The data points of the problem represent the statistics of 7073 customers of a fictional Telco company. The concerned label is whether the customer is going to cancel the service or not (1 for yes, 0 for no). The features include services that each customer has signed up for, customer account information, and demographic info about customers.

The dataset was collected from Kaggle: https://www.kaggle.com/datasets/blastchar/telco-customer-churn A newer version of the dataset is also available from IBM: https://community.ibm.com/community/user/businessanalytics/blogs/steven-macko/2019/07/11/telco-customer-churn-1113

I performed feature engineering, feature selection and built two models for comparision using Logistic Regression and Decision Tree Classifier.

The final model using Logistic Regression reaches an accuracy of approximately 79% on the test set. It is compared with another model using Decision Tree Classifier, which only yields an accuracy of 73%.

To run this project, the following libraries should be installed:

• pandas
• matplotlib
• Sklearn
• matplotlib
• seaborn