AI Stock Advisor Web Application

A modern, elegant web application for AI-powered stock recommendations with a sophisticated frontend and Flask backend.

Requirements

Python Dependencies

Create a requirements.txt file with the following:

flask>=2.0.0
flask-cors>=3.0.0
pandas>=1.5.0
numpy>=1.20.0
scikit-learn>=1.0.0
lightgbm>=3.0.0
joblib>=1.0.0
yfinance>=0.1.0

Installation

# Install dependencies
pip install -r requirements.txt

# Or install individually:
pip install flask flask-cors pandas numpy scikit-learn lightgbm joblib yfinance

File Structure

Your project directory should look like this:

ai-stock-advisor/
├── server.py                          # Flask backend server
├── index.html                         # Frontend web interface  
├── requirements.txt                   # Python dependencies
├── feature_engineered_data.csv        # Your processed data
├── short_term_model.joblib            # 30-day model
├── medium_term_model.joblib           # 90-day model
├── long_term_model.joblib             # 365-day model
└── README.md                          # This file

Quick Start

1. Ensure your models and data are ready:

   • Run your existing Jupyter notebook to generate the required files
   • Make sure you have: feature_engineered_data.csv and the three .joblib model files

2. Save the frontend:

   • Save the HTML code as index.html in your project directory

3. Save the backend:

   • Save the Python server code as server.py

4. Install dependencies:

   pip install -r requirements.txt

5. Start the server:

   python server.py

6. Open your browser:

   • Navigate to http://localhost:5000
   • The modern web interface will load automatically

API Endpoints

POST /api/recommendations

Get stock recommendations based on user preferences.

Request Body:

{
    "risk_level": "Medium",
    "time_horizon": "medium", 
    "num_stocks": 10
}

Response:

{
    "success": true,
    "recommendations": [
        {
            "symbol": "RELIANCE.NS",
            "price": 2456.75,
            "volatility": 0.0234,
            "volume": 1234567,
            "rsi": 65.4,
            "macd": 12.3
        }
    ],
    "parameters": {
        "risk_level": "Medium",
        "time_horizon": "medium",
        "num_stocks": 10
    }
}

GET /api/chart/{symbol}

Get historical price data for charting.

Response:

{
    "symbol": "RELIANCE.NS",
    "dates": ["2023-01-01", "2023-01-02", ...],
    "close": [2400, 2405, ...],
    "sma_20": [2390, 2392, ...],
    "sma_50": [2385, 2387, ...],
    "volume": [1000000, 1100000, ...]
}

GET /api/status

Check server status and loaded data.

Features

Frontend

• Modern Design: Glass-morphism effects, gradients, and smooth animations
• Responsive Layout: Works perfectly on desktop, tablet, and mobile
• Interactive Controls: Intuitive sliders and dropdowns for user preferences
• Real-time Charts: Interactive price charts using Chart.js
• Loading States: Professional loading indicators and transitions
• Error Handling: Graceful error messages and troubleshooting guides

Backend

• RESTful API: Clean, well-documented endpoints
• Model Integration: Seamlessly loads and uses your trained models
• Data Processing: Handles feature engineering and predictions
• Error Handling: Comprehensive error catching and logging
• CORS Support: Enables frontend-backend communication

Key Improvements over Streamlit

1. Performance: Much faster loading and response times
2. Customization: Full control over UI/UX design
3. Scalability: Can handle multiple concurrent users
4. Professional Look: Modern, polished interface
5. API-First: Easy to integrate with mobile apps or other services

Troubleshooting

Common Issues

1. "Model files not found"

   • Ensure you've run your Jupyter notebook to generate the .joblib files
   • Check that all three model files are in the same directory as server.py

2. "Feature-engineered data not found"

   • Run the feature engineering script from your notebook
   • Verify feature_engineered_data.csv exists and has data

3. "Connection refused" or API errors

   • Make sure the Flask server is running: python server.py
   • Check that no other service is using port 5000
   • Try accessing http://localhost:5000/api/status directly

4. Charts not loading

   • Ensure the backend is returning chart data
   • Check browser console for JavaScript errors
   • Verify the stock symbol exists in your dataset

Development Tips

• Debug Mode: The server runs in debug mode by default for development
• Logging: Check the console output for detailed request/response information
• Data Validation: The server includes comprehensive input validation
• Model Performance: Each prediction includes performance metrics

Production Deployment

For production use:

1. Set debug=False in the Flask app
2. Use a production WSGI server like Gunicorn
3. Add proper authentication if needed
4. Set up HTTPS with SSL certificates
5. Configure CORS for your production domain

Security Considerations

• The current setup is for development/demo purposes
• Add input validation and rate limiting for production
• Consider authentication for sensitive financial data
• Use HTTPS in production environments

Next Steps

1. Add More Features: Portfolio optimization, backtesting, alerts
2. Mobile App: Use the API to create a mobile application
3. Real-time Data: Integrate live market data feeds
4. User Accounts: Add user registration and portfolio tracking
5. Advanced Charts: Add more technical indicators and chart types