🏦 Loan Approval Predictor (ML + Flask + Authentication)

A machine-learning-based Loan Approval System with user authentication,
Risk Score prediction, and an interactive dashboard built using Flask.

---

📌 Overview

This project leverages machine learning models trained on a synthetic dataset to:

• Predict a Risk Score for loan applicants using LGBM Regressor
• Predict whether a Loan will be Approved/Rejected using Random Forest Classifier
• Display prediction history & interactive charts for insights
• Provide a secure login/signup system for users

The web interface is built using Flask, HTML, CSS, and includes user authentication for restricted access.

---

🔐 User Authentication (NEW)

A complete authentication system is added:

• Login page
• Signup page
• Session-based login
• Protected dashboard (index.html)
• Users stored in users.json

Only logged-in users can access:

• Prediction Form
• Dashboard
• Results

🔑 Authentication Screenshots

🔵 Login Page

🟣 Signup Page

🟢 Dashboard

---

📁 Project Structure

Loan-Approval-Predictor/
│
├── models/                   # Saved ML models and scaler
│   ├── clf_model.pkl
│   ├── reg_model.pkl
│   └── scaler.pkl
│
├──static/                   # Static assets (CSS, images)
│   ├── style.css
│   ├── login.png
│   ├── signup.png
│   ├── dashboard.png
|
├── templates/                # Flask HTML templates
│   ├── index.html
│   ├── result.html
│   ├── login.html            
│   ├── signup.html           
│
├── app.py                    # Flask backend
├── .gitignore
├── requirements.txt
├── Loan.csv                  # Dataset
├── EDA.ipynb                 # Data analysis & feature engineering
└── model.ipynb               # Model training & selection

🔧 Data & Modeling

All preprocessing, feature engineering, and model experimentation are documented in:

• EDA.ipynb → Cleaning, visualization, correlation, feature engineering
• model.ipynb → Model training & evaluation

✔ Final ML Pipeline

• Features standardized and numeric columns log-transformed (log1p)
• LGBM Regressor → Predicts Risk Score
• Random Forest Classifier → Predicts Loan Approval based on Risk Score
• Saved models:
  • reg_model.pkl
  • clf_model.pkl
  • scaler.pkl

---

🚀 Deployment

The Flask application provides:

📂 Templates

• index.html → Input form, progress bar, visual dashboard
• result.html → Shows score, approval result, pie charts
• login.html → Login UI
• signup.html → Signup UI

📁 Backend

• app.py
  • Routes for login, signup, prediction
  • Session authentication
  • Returns history + chart data as JSON

🎨 Static

• style.css
• login.png, signup.png, dashboard.png (screenshots)

---

🧠 Results Summary

• Risk Score Prediction: LGBM Regressor with standardized features
• Loan Approval Prediction: Random Forest Classifier
• Dashboard: Progress bar, history table, pie charts
• Authenticated Access: Only logged-in users can use the tool

---

💻 Future Improvements

• Save prediction history in a database.
• Add more visual analytics for loan trends.
• Deploy on Heroku or AWS for live access.
• Add password hashing for higher security

---

💻 How to Use This Project

1. Clone the Repository

git clone https://github.com/your-username/loan-approval-predictor.git
cd loan-approval-predictor

2. Create a Virtual Environment

python -m venv venv
# On Windows:
venv\Scripts\activate
# On macOS/Linux:
source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Run the App

python app.py

Then open your browser and go to: 📍 http://127.0.0.1:5000

---