Credit Risk Modelling App

Overview

This project implements an end-to-end machine learning system for predicting credit risk using customer financial data. The model is trained on structured data and deployed via a Streamlit application for real-time inference.

The workflow separates:

• Model development (notebook)
• Model usage (app)

---

Project Structure

.
├── data/
│   └── german_credit_data.csv
│
├── models/                        # Generated locally (not tracked in Git)
│   ├── model.pkl
│   ├── encoders.pkl
│   └── scaler.pkl
│
├── notebooks/
│   └── credit_risk.ipynb         # Model training and experimentation
│
├── src/
│   └── app.py                    # Streamlit application
│
├── requirements.txt
└── README.md

---

Important Note

The trained model and preprocessing artifacts (.pkl files) are not included in the repository.

They must be generated before running the application.

---

Setup & Usage

Step 1: Train Model (Required)

Run the notebook:

jupyter notebook notebooks/credit_risk.ipynb

Execute all cells to generate:

• models/model.pkl
• models/encoders.pkl
• models/scaler.pkl

---

Step 2: Run Application

streamlit run src/app.py

This will launch a web interface for predicting credit risk.

---

Machine Learning Pipeline

Data Preprocessing

• Encoding categorical variables
• Feature scaling

Model Training

• Trained classification model (e.g., Random Forest)
• Model selected based on performance

Deployment

• Streamlit-based UI for real-time predictions

---

Author

Mainak Jana