🖥️ Lending Risk Assessment Platform (LDRA)

An intelligent, cloud-native banking and risk-assessment dashboard. LDRA replaces manual spreadsheet analysis by bridging a modern React web app with a Python Machine Learning microservice to instantly predict loan default probabilities in real-time.

---

✨ Key Features

• 🧠 Predictive Analytics: Uses a trained scikit-learn Random Forest to output a live Default Probability and Risk Level (Low, Medium, High).
• 📊 Real-time Dashboard: Built iteratively with Recharts providing live macro-metrics on Approval Rates, Debt-to-Income distributions, and aggregate portfolio values modeled for Indian limits.
• 🔒 Secure Banking Standards: Requires strict password architecture and protects routes with hardened JSON Web Tokens (JWT).

---

🏗 Architecture & Stack

1. Frontend (Vite + React)

• Hosted on Vercel for lightning-fast Edge-CDN delivery.
• Uses wouter for lightweight routing, @tanstack/react-query for server synchronization, and lucide-react for scalable UI.

2. Core Backend API (Node.js + Express)

• Hosted on Render, orchestrating records into MongoDB Atlas.
• Responsible for authentication, fetching dynamic chart mappings, and proxying data securely to the Python ML Engine.

3. ML Risk Microservice (Python + Flask)

• Hosted on Render utilizing gunicorn.
• Evaluates borrower parameters (Credit Score, Term, Income, Age, Debt Constraints) against a pre-trained .pkl model to catch bad loans before they happen.

---

☁️ Deployment

1. The React UI can be directly connected to Vercel.
2. The Node Backend can be deployed as a Render Web Service (Command: npm start).
3. The Flask ML Server can be deployed as a Render Web Service (Build Command: pip install -r requirements.txt, Start Command: gunicorn app:app).