Credit Risk Modeling with German Credit Data

This project implements an end-to-end credit risk modeling pipeline using the classic German Credit dataset. It follows the IFRS 9 / Basel framework, covering:

• PD (Probability of Default) modeling with logistic regression + validation and calibration
• LGD (Loss Given Default) simulation and modeling with Random Forests
• EAD (Exposure at Default) simulation based on credit utilization assumptions
• ECL (Expected Credit Loss) calculation with portfolio summaries and segment analysis
• Stress testing under adverse economic scenarios

The workflow includes EDA, model training, visualization, and portfolio risk analysis.

---

Jupyter Notebook

For a full walkthrough with code, outputs, and visualizations, see the Jupyter Notebook Credit_Risk_Analysis.ipynb

Run the notebook online (no setup required):