Customer Churn Prediction

Project Information

View on GitHub
  • Category: Predictive Analytics
  • Technologies: Scikit-learn XGBoost Flask Plotly
  • Project Date: July 2024
  • Dataset: Telco Customer Churn
  • Accuracy: 87.5% (cross-validated)

Customer Churn Prediction

A machine learning model that predicts customer churn for companies, helping businesses identify at-risk customers and implement retention strategies.

Technical Implementation

  • Multiple classification algorithms compared (Logistic Regression, Random Forest, XGBoost)
  • Feature importance analysis using SHAP values
  • Handling of class imbalance with SMOTE technique
  • Hyperparameter optimization using GridSearchCV

Key Features

  • Early identification of at-risk customers
  • Customer segmentation based on churn probability
  • Interactive dashboard for business stakeholders
  • Exportable reports with retention recommendations
Churn Prediction Dashboard
Feature Importance Analysis
Customer Segmentation

Video Demonstration

See the Customer-Churn-Prediction Identification system in action

This demonstration shows the complete workflow of the Customer-Churn-Prediction system.