Credit Risk Prediction API

A FastAPI-based backend service that predicts loan approval risk using a trained machine learning model. The system accepts applicant financial data, calculates the probability of default, and returns an approval or rejection decision while storing predictions in a database.

This project demonstrates how to integrate machine learning inference with a production-style backend system using FastAPI, SQLAlchemy, and Docker.

---

Live API

• Base URL: https://credit-risk-api-wpam.onrender.com
• Swagger Docs: https://credit-risk-api-wpam.onrender.com/docs

---

Features

• Machine learning–based credit risk prediction
• REST API built with FastAPI
• SQLAlchemy database integration for storing prediction records
• Probability-based loan approval decision system
• Docker support for containerized deployment
• Retrieval of previous prediction records through API
• Deployed on Render with PostgreSQL

---

Tech Stack

Layer	Technology
Backend	Python, FastAPI
Machine Learning	Scikit-learn (serialized model artifact)
Database	PostgreSQL (Render), SQLAlchemy ORM
Deployment	Docker, Render (Cloud)

---

Project Structure

credit-risk-api
│
├── app
│   ├── main.py            # FastAPI application and endpoints
│   ├── config.py           # Configuration (approval threshold)
│   ├── inference.py        # ML model loading and prediction
│   ├── schemas.py          # Pydantic request/response models
│   │
│   └── db
│       ├── database.py     # Database connection and session
│       ├── models.py       # SQLAlchemy table definitions
│       ├── crud.py         # Database operations
│
├── credit_risk_artifacts.pkl   # Trained ML model bundle
├── create_tables.py
├── requirements.txt
├── Dockerfile
└── .env.example

---

System Workflow

Client Request
      ↓
FastAPI Backend (validates input)
      ↓
ML Model (predicts risk probability)
      ↓
Decision Engine (approved / rejected)
      ↓
PostgreSQL (stores prediction record)
      ↓
JSON Response → Client

1. Client sends loan applicant details to the /predict endpoint.
2. FastAPI backend receives and validates the request.
3. The trained ML model loads from credit_risk_artifacts.pkl.
4. The model calculates risk probability.
5. Backend compares probability with a configured approval threshold.
6. Decision is generated: approved or rejected.
7. Prediction data is stored in the database.
8. Response is returned to the client.

---

API Endpoints

Health Check

GET /

Response:

{
  "message": "credit risk service running"
}

---

Predict Credit Risk

POST /predict

Request:

{
  "no_of_dependents": 2,
  "income_annum": 450000,
  "loan_amount": 120000,
  "loan_term": 12,
  "cibil_score": 720,
  "bank_asset_value": 300000
}

Response:

{
  "risk_probability": 0.27,
  "decision": "approved"
}

---

Get Previous Predictions

GET /predictions?limit=10&offset=0

Parameter	Description
limit	Number of records to return
offset	Pagination offset

---

Live Example

curl -X POST https://credit-risk-api-wpam.onrender.com/predict \
  -H "Content-Type: application/json" \
  -d '{
    "no_of_dependents": 2,
    "income_annum": 450000,
    "loan_amount": 120000,
    "loan_term": 12,
    "cibil_score": 720,
    "bank_asset_value": 300000
  }'

{
  "risk_probability": 0.27,
  "decision": "approved"
}

---

Running Locally

# Clone the repository
git clone https://github.com/DSurya11/credit-risk-api.git
cd credit-risk-api

# Create and activate virtual environment
python -m venv venv
venv\Scripts\activate        # Windows
# source venv/bin/activate   # macOS/Linux

# Install dependencies
pip install -r requirements.txt

# Configure environment
copy .env.example .env
# Edit .env with your database credentials

# Run the API
uvicorn app.main:app --reload

Server: http://127.0.0.1:8000 Swagger docs: http://127.0.0.1:8000/docs

---

Docker Setup

# Build
docker build -t credit-risk-api .

# Run
docker run -p 8000:8000 credit-risk-api

---

Future Improvements

• Authentication and authorization
• Model monitoring and retraining pipeline
• API rate limiting
• Feature store for model inputs
• Logging and observability

---

Author

Surya GitHub: DSurya11