💳 Credit Scoring System
Predict Creditworthiness Using Machine Learning, Streamlit Web App, and Flask API
This project provides a complete credit scoring solution that predicts whether a person is creditworthy based on personal and financial information. It includes a training pipeline, a web app interface, a REST API, and feature importance visualization.
🏗 Project Overview Features:
End-to-end Machine Learning pipeline:
Handles numeric and categorical features.
Scales numeric columns and encodes categorical columns automatically.
Random Forest Classifier:
Balanced to handle imbalanced classes (good/bad credit).
Hyperparameter tuning with GridSearchCV.
Deployment-ready:
Streamlit Web App for interactive user predictions.
Flask REST API for integration with other systems.
Explainability:
Top features influencing predictions are visualized with a bar chart.
✅ Key Benefits
Quickly assess creditworthiness for loan approvals.
Understand which features most influence decisions (feature importance).
Easy integration into web apps or existing financial systems.
Handles unseen or unknown categories safely.
CreditScoring/ │ ├─ data/ │ └─ german_credit_data.csv # Dataset │ ├─ models/ │ └─ credit_model.pkl # Trained ML pipeline │ ├─ train_model.py # Model training script ├─ streamlit_app.py # Streamlit interactive web app ├─ app.py # Flask REST API ├─ requirements.txt # Python dependencies └─ README.md # Project documentation