💳 Credit Default Risk Prediction

A production-ready machine learning web application to predict whether a credit card client will default in the next month.

This project uses a real-world dataset from the UCI Machine Learning Repository, combined with feature engineering, model tuning, and SHAP explainability — all deployed via Streamlit.

---

Demo

Project Structure

credit_scoring_model/
├── app.py                 # main Streamlit app
├── requirements.txt       # python dependencies
├── models/                # trained model (.pkl)
├── data/                  
├── notebook/              # notebook for training & EDA

Problem Statement

Predict whether a customer will default on their credit card payment next month based on demographic data, bill/payment history, and credit behavior.

Dataset Description

• Source: Kaggle Dataset
• Rows: 30,000
• Columns: 24 original + 5 engineered features

Input Features:

Column	Description
LIMIT_BAL	Credit limit (NT dollars)
SEX	Gender (1 = Male, 2 = Female)
EDUCATION	Education level (1 = Graduate, 2 = University, ...)
MARRIAGE	Marital status
AGE	Age in years
PAY_0 to PAY_6	Repayment status (last 6 months)
BILL_AMT1 to BILL_AMT6	Monthly bill amounts
PAY_AMT1 to PAY_AMT6	Monthly payment amounts

Engineered Features:

• TOTAL_PAY_AMT: Total amount paid in 6 months
• TOTAL_BILL_AMT: Total bill in 6 months
• NUM_LATE_PAYMENTS: Number of late payments
• MAX_DELAY: Longest delay (in months)
• LONGEST_LATE_STREAK: Longest continuous months of delay

Tech Stack

• Language: Python 3.12
• ML Model: XGBoost Classifier
• Tuning: GridSearchCV
• Imbalance Handling: SMOTE
• Explainability: SHAP (Waterfall plot)
• Deployment: Streamlit

---

How to Run Locally

1. Clone the repository

   git clone https://github.com/your-username/credit-scoring-model.git
   cd credit-scoring-model

2. Create virtual environment

   python -m venv venv
   venv\Scripts\activate  # Windows
   source venv/bin/activate  # Mac/Linux

3. Install dependencies

   pip install -r requirements.txt

4. Run the app

   streamlit run app.py

---

Model Performance

Metric	Score
ROC-AUC	0.7815
Accuracy	82%
Tuning	GridSearchCV
Explainability	SHAP waterfall

SHAP Visualization

We use SHAP to explain predictions for each customer:

Future Improvements

• Add login/auth for secure access
• Store prediction history to database
• Train model incrementally with real-time data

---

🧑‍💻 Author

Đạt Đình
2nd Year @ UET, Data Engineering Track
Email: dinhdatnguyen0710@example.com

📜 License

MIT License — Free to use and modify

🙌 Acknowledgments

• UCI Credit Card Dataset
• Streamlit community
• SHAP and XGBoost developers