Bank Customer Churn Predictor 🏦

Overview

An end-to-end machine learning application that predicts customer churn probability and generates personalized engagement strategies using ML models and LLMs.

🔥 Features

• Multi-Model Prediction: Ensemble of 9 different ML models including XGBoost, Random Forest, and SVM
• Interactive Dashboard: Built with Streamlit for easy data input and visualization
• AI-Driven Engagement: Personalized customer retention strategies using Groq's LLM
• Advanced ML Techniques:
  • SMOTE for handling imbalanced data
  • Feature engineering for improved accuracy
  • Voting classifier for ensemble predictions

🚀 Live Demo

Try out the live application here: Bank Churn Predictor

💻 Tech Stack

• Python 3.12
• Streamlit
• Scikit-learn
• XGBoost
• Pandas
• NumPy
• Groq API
• Plotly

📊 Model Performance

• XGBoost with SMOTE: 86% accuracy
• Feature-engineered model: 88% accuracy
• Voting Classifier: 87% accuracy

🛠️ Installation & Setup

1. Clone the repository

git clone https://github.com/yourusername/bank-churn-predictor.git
cd bank-churn-predictor

2. Install dependencies

pip install -r requirements.txt

3. Set up environment variables

# Create a .env file in the root directory and add:
GROQ_API_KEY=your_api_key_here

4. Run the application

streamlit run main.py

Prerequisites

• Python 3.11 or higher
• Groq API key (get it from Groq's website)
• Git

Note

Make sure all model files (.pkl) are in the models/ directory and the dataset (churn.csv) is in the data/ directory before running the application.

📁 Project Structure

bank-churn-predictor/
├── main.py                     # Main Streamlit application
├── utils.py                    # Utility functions and visualizations
├── requirements.txt            # Project dependencies
├── .env                        # Environment variables (Groq API key)
├── models/                     # Trained ML models
│   ├── xgb_model.pkl          # XGBoost model
│   ├── nb_model.pkl           # Naive Bayes model
│   ├── rf_model.pkl           # Random Forest model
│   ├── dt_model.pkl           # Decision Tree model
│   ├── svm_model.pkl          # SVM model
│   ├── knn_model.pkl          # K-Nearest Neighbors model
│   ├── voting_clf.pkl         # Voting Classifier model
│   ├── xgboost-SMOTE.pkl      # XGBoost with SMOTE
│   └── xgboost-featureEngineered.pkl  # Feature-engineered XGBoost
├── data/                       # Dataset directory
│   └── churn.csv              # Bank customer dataset
├── notebooks/                  # Jupyter notebooks
│   └── model_training.ipynb    # Model training and analysis
└── README.md                   # Project documentation

🔮 Future Improvements

• Retrain models with different feature engineering approaches
• Implement GradientBoostingClassifier and StackingClassifier
• Explore different LLM options and prompting techniques
• Deploy ML models as separate API endpoints
• Test models on different churn datasets

📝 Blog Post

Read about the development process and technical details in my blog post: How I Built a System to Predict and Prevent Bank Customer Churn Using ML, LLMs, and Streamlit

📫 Contact

• LinkedIn: Alex Wang
• Twitter: @imalexwang
• Blog: solo diaries