Loan Eligibility Prediction Web App

📁 Description

This project is a full-stack web application that predicts a user's loan eligibility based on their financial and personal details. The app is built with a React frontend, Flask backend, and a machine learning model trained on historical loan data.

📁 Project Structure

loan-eligibility-app/
│
├── frontend/          # React.js application for user interaction
├── model/             # Trained machine learning model (.pkl file)
├── train.csv          # Dataset used to train the model
├── README.md          # Project documentation
├── model/             # Trained machine learning model (.pkl file)
  └── LoanPrediction.ipynb  # Notebook to preprocess, train, and export the model

🔧 Backend

• Built with Flask
• Accepts POST requests at /predict with 11 user inputs
• Preprocesses data to match training pipeline
• Loads the trained model (logistic_model.pkl) and returns predictions
• Uses flask-cors to support requests from the React frontend

Dependencies:

• Flask
• numpy
• pickle
• flask-cors

💻 Frontend

• Built with React.js (Vite or Create React App)
• Accepts user input through form fields
• Sends data to backend /predict route via fetch
• Displays model prediction as "Approved" or "Rejected"
• Styled using Tailwind CSS with subtle animations

🧠 Model

• Trained using RandomForestClassifier
• Uses 11 engineered features like Gender, Income, Credit History, etc.
• Trained and saved using TrainAndSaveModel.ipynb
• Model is serialized using pickle and stored at model/logistic_model.pkl

📈 Algorithms Used

1. RandomForestClassifier

   • An ensemble learning method that builds multiple decision trees and merges their results to improve accuracy and reduce overfitting.

2. LogisticRegression (Optional)

   • A linear classification algorithm used for binary outcomes (like "Approved" vs "Rejected").

You can switch between algorithms in the training notebook by changing the classifier line.

⚡ How to Run the Project

Prerequisites:

• Python 3.x
• Node.js & npm
• pip (Python package installer)

1. Clone the Repository

git clone <your-repo-url>
cd loan-eligibility-app

2. Backend Setup

cd backend
pip install -r requirements.txt
python app.py

3. Frontend Setup

cd frontend
npm install
npm run dev

4. Train the Model (Optional if model is not already saved)

Open TrainAndSaveModel.ipynb in Jupyter or VS Code
Run all cells to save the model to /model/logistic_model.pkl

Make sure your Flask backend is running at http://localhost:5000 when testing the frontend.

✅ Result

• Provides instant loan eligibility predictions
• Successfully integrates ML with a responsive UI
• Handles edge cases and invalid inputs with graceful messages

📚 Learning

• Hands-on practice integrating ML + Flask + React
• How to preprocess and encode categorical data
• Deploying full-stack ML apps locally
• Managing API-CORS issues and model versioning

🚀 Future Improvements

• Deploy to cloud (Render, Vercel, Railway, etc.)
• Add login/authentication support
• Save prediction history to a database (e.g. SQLite or MongoDB)
• Add more advanced ML models or tuning (GridSearchCV)
• Improve UI with charts or personalized feedback

---

Built with ❤️ using Python, React, and Machine Learning