About This Platform

Customer Behavior Prediction Platform – Vantara Retail Solutions

Project Overview

This platform helps the Marketing and Retention teams identify customers who are at risk of churning. It uses real transaction data from an online retail business and applies machine learning models to predict churn probability, estimate customer value, and segment customers based on their behavior.

Key Capabilities

Churn Prediction

Predicts the probability that a customer will stop purchasing in the near future.

Customer Lifetime Value

Estimates the historical and potential value of each customer.

Customer Segmentation

Groups customers into meaningful segments for targeted marketing.

Model Explainability

Uses SHAP and LIME to explain why a customer is predicted to churn.

Technology Stack

Data & Features: Python, Pandas, NumPy

Machine Learning: Scikit-learn, XGBoost, LightGBM

Deep Learning: TensorFlow (ANN, LSTM, Autoencoder)

Explainability: SHAP, LIME

Backend: FastAPI

Frontend: HTML, CSS, JavaScript

Project Status

This is an internship project developed for Vantara Retail Solutions. The platform currently supports customer lookup, churn risk scoring, and a high-risk leaderboard. Further improvements such as Docker deployment and additional dashboard features are planned.