scorecard-toolkit 2.0.0

A professional and modern toolset for scorecard modeling, fully compatible with scikit-learn

ScoreCardModel

ScoreCardModel is a professional and modern toolset for scorecard modeling, fully compatible with scikit-learn. It is designed for credit risk analysts and data scientists who need to build transparent, regulator-friendly scoring models with ease.

🚀 Key Features

• 🛠 Scikit-Learn Compatible: BinningTransformer, WOETransformer, and ScoreCardTransformer work in any Pipeline or GridSearchCV.
• 📊 Rich Analytics: 18+ plot types (KS, ROC, CAP, Lift, Calibration, PSI, etc.) for comprehensive model evaluation.
• 📝 Automated Reporting: Generate professional Markdown or Excel reports with one function call.
• 🔄 5 WOE Methods: Standard, Adjusted (Laplace), Empirical Logit, Signed, and Weighted Weight of Evidence.
• 🎮 Interactive Dashboard: A Jupyter-based what-if widget for real-time scorecard testing.
• 🏢 Industry Standard: Built-in support for PDO (Points to Double Odds) and Base-Odds scaling.

📸 Visual Gallery

KS Curve	ROC Curve	CAP Curve

Score Distribution	Scorecard Waterfall	IV Summary

See all 12+ visualizations →

📦 Installation

pip install scorecard-toolkit

⚡ Quick Start

from ScoreCardModel import ScoreCardWrapper

# Initialize and fit
sc = ScoreCardWrapper(binning_strategy='quantile', base_points=600, pdo=20)
sc.fit(X_train, y_train)

# Predict scores and export scorecard
scores = sc.predict(X_test)
card = sc.export_scorecard()
print(card.head(10))

📚 Documentation

Visit scorecardmodel.readthedocs.io for the full documentation, including:

• Installation Guide
• Quickstart Guide
• Full Examples & Visualizations
• WOE In-Depth Guide
• API Reference

🤝 Contributing

We welcome contributions! Please see our Contributing Guidelines for more details.

📄 License

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