UgeneTettey/African_Telecom_Company

Capstone project under Part-Time Digital Africa Data Science Training at Blossom Academy. TELECOM CUSTOMER CHURN PREDICTION Overview In the competitive telecommunications industry, customer churn - where customers become inactive and stop purchasing services - is a significant challenge. This project aims to develop a machine learning model to predict the likelihood of customer churn, helping the company to proactively retain its customer base. We also aim to access the likelihood of customers becoming inactive and discontinuing their purchase of airtime and data for a period of 90 days. By leveraging customer data, the model will identify customers at risk of churning, enabling the company to implement targeted retention strategies and interventions. Project Roadmap This project was a collaboration among dedicated minds who contributed to its success. Preprocessing steps included tenure mapping, label encoding, and class balancing (using downsampling). Featured models include: 1. Logistic Regression (Baseline Model) 2. Decision Tree Classifier 3. RandomForest Classifier 4. Gradient Boosting Classifier 5. XGBoost Classifier (Adopted Model) Model was deployed utilising steamlit. The link to the deployed app is available in the "About" section of this repository.