Customer Churn Prediction And Retention Strategy Optimization For Subscription-Based Services Using Behavioural Data Analytics And Machine Learning Models

This project presents a user-friendly Customer Churn Prediction and Retention Dashboard developed using Streamlit. The tool enables business users and analysts to upload customer datasets in CSV format, preprocess the data, and apply multiple machine learning models to predict customer churn. Supported models include Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, Gradient Boosting, XGBoost, LightGBM, and CatBoost. The dataset is automatically encoded and scaled for compatibility with the models. Once trained, the selected model predicts customer churn and evaluates performance using metrics such as accuracy score, classification report, and confusion matrix. The dashboard also computes churn probabilities and recommends retention actions for high-risk customers (e.g., offering incentives to those with a churn probability > 0.7). It includes visualizations such as churn distribution, retention action breakdown, and feature importance for better interpretability. This approach not only aids in understanding customer behavior but also supports proactive decision-making to improve retention. Users can download the enriched dataset with churn probabilities and retention strategies for further business use.