Credit Risk and Loan Default Prediction System Machine learning system for credit risk assessment using German Credit and Loan Approval datasets. Implements multiple classification algorithms with comprehensive exploratory data analysis, feature engineering, and model comparison visualizations. 📊 Project Overview This project analyzes credit risk using two distinct datasets:

German Credit Dataset: Assesses individual creditworthiness based on financial and personal attributes Loan Approval Dataset: Predicts loan approval/rejection based on applicant demographics and financial history

The system trains and compares three machine learning models, performs extensive feature engineering, and generates detailed visualizations for model interpretation. 🎯 Features

Multiple ML Models: Logistic Regression, Random Forest, Gradient Boosting Automated EDA: Comprehensive exploratory data analysis with 9+ visualization plots per dataset Feature Engineering: Custom risk scores, income ratios, and composite indicators Model Comparison: Side-by-side performance metrics and ROC curves Prediction Analysis: Confusion matrices, probability distributions, and error analysis Feature Importance: Visual ranking of most influential predictors Cross-Validation: 5-fold stratified CV for robust performance estimates Automated Reporting: JSON export of all results and conclusions

🚀 Getting Started Prerequisites bashPython 3.7+ Installation

Clone the repository:

bashgit clone https://github.com/yourusername/credit-risk-analysis.git cd credit-risk-analysis

Install required packages:

bashpip install pandas numpy matplotlib seaborn scikit-learn Usage Run the complete analysis: bashpython credit_risk_analysis.py The script will:

Load or generate sample datasets Perform exploratory data analysis Train multiple ML models Generate visualizations Create a final report

📁 Project Structure credit-risk-analysis/ │ ├── credit_risk_analysis.py # Main analysis script ├── README.md # This file │ ├── data/ # (Optional) Place your datasets here │ ├── credit_risk.csv │ └── loan_train.csv │ └── output/ # Generated files ├── german_credit_eda.png ├── loan_approval_eda.png ├── german_model_comparison.png ├── loan_model_comparison.png ├── german_prediction_analysis.png ├── loan_prediction_analysis.png ├── dataset_comparison.png ├── german_feature_importance.png ├── loan_feature_importance.png ├── project_report.json ├── german_credit_model.pkl └── loan_approval_model.pkl 📊 Datasets German Credit Dataset

Size: 1,000 samples Features: 17 attributes including checking balance, credit history, loan purpose, employment duration Target: Binary classification (default/no default)

Loan Approval Dataset

Size: 600+ samples Features: 12 attributes including income, loan amount, credit history, property area Target: Binary classification (approved/rejected)

🔬 Methodology

1. Data Preprocessing

Missing value imputation (median for numeric, mode for categorical) Label encoding for categorical variables Feature scaling using StandardScaler

2. Feature Engineering

German Credit: Risk scores for checking balance, savings, employment Loan Approval: Total income, loan-to-income ratio, monthly EMI calculations Composite risk indicators

3. Model Training Three models trained with optimized hyperparameters:

Logistic Regression: L2 regularization, balanced class weights Random Forest: 50-100 estimators, max depth constraints Gradient Boosting: 50-100 estimators, learning rate optimization

4. Evaluation Metrics

Accuracy Precision Recall F1-Score ROC-AUC (primary metric) 5-fold cross-validation scores

📈 Results German Credit Dataset

Best Model: Random Forest AUC: ~0.75-0.80 Key Features: Checking balance, credit history, loan duration

Loan Approval Dataset

Best Model: Gradient Boosting / Random Forest AUC: ~0.70-0.85 Key Features: Credit history, total income, loan amount

🖼️ Visualizations The system generates 9 comprehensive visualization files:

EDA Plots: Distribution analysis, default rates by categories Model Comparison: Bar charts comparing accuracy, AUC, F1-scores ROC Curves: Multi-model performance comparison Prediction Analysis: Confusion matrices, probability distributions Feature Importance: Top 10 most influential features Dataset Comparison: Cross-dataset model performance

🛠️ Configuration Custom Dataset Usage Place your CSV files in the project directory with names:

credit_risk.csv or german_credit.csv (German Credit data) loan_train.csv or train.csv (Loan Approval data)

Model Hyperparameters Modify in the train_german_models() and train_loan_models() methods: pythonmodels = { 'Random Forest': RandomForestClassifier( n_estimators=100, # Adjust here max_depth=10, # Adjust here random_state=42 ) } 📝 Output Files

PNG Images: High-resolution (300 DPI) visualizations JSON Report: Complete analysis summary with metrics PKL Models: Serialized best models for deployment

🤝 Contributing Contributions are welcome! Please feel free to submit a Pull Request.

Fork the project Create your feature branch (git checkout -b feature/AmazingFeature) Commit your changes (git commit -m 'Add some AmazingFeature') Push to the branch (git push origin feature/AmazingFeature) Open a Pull Request

📄 License This project is licensed under the MIT License - see the LICENSE file for details. 👥 Authors

Andy Zhu (AZ455) Jacob Cheng (JC3045)

Course: CS 439: Introduction to Data Science 🙏 Acknowledgments

German Credit Dataset: UCI Machine Learning Repository Loan Approval Dataset: Analytics Vidhya scikit-learn documentation and community

📧 Contact For questions or feedback, please open an issue on GitHub.