An end-to-end AI-powered web application that predicts customer churn and provides clear, data-driven explanations using Machine Learning, SHAP, and LLM-based insights.
- Predicts whether a customer is likely to churn
- Displays churn probability with a progress indicator
- Uses LLM (Groq API) to explain predictions in simple language
- Ensures explanations are consistent and data-aware
- Visualizes feature impact on prediction
- Highlights top contributing factors
- Ask follow-up questions about the prediction
- Context-aware responses for better understanding
- Customer churn distribution
- Age distribution
- Balance vs churn analysis
- Product usage impact
- Frontend: Streamlit
- Backend: Python
- ML Model: Random Forest Classifier
- Explainability: SHAP
- AI Integration: Groq (LLM API)
- Data Processing: Pandas, NumPy
- Visualization: Matplotlib
├── main.py
├── ai_explainer.py
├── model.pkl
├── Churn_Modelling.csv
├── requirements.txt