Stock Market Predictor

A full-stack machine learning application that predicts next-day stock price movements using a Random Forest classifier.

Features

• 🤖 Machine Learning: Random Forest model with 51.88% test accuracy
• 🔮 Stock Predictions: Predicts UP/DOWN movement for next trading day
• 📊 Technical Analysis: Uses 30+ technical indicators (RSI, MACD, Bollinger Bands, etc.)
• 📈 Interactive Charts: 3-month historical price visualization
• 🌐 REST API: Flask backend with comprehensive endpoints
• ⚛️ Modern UI: React frontend with Tailwind CSS
• 🚀 Production Ready: Configured for Railway (backend) and Vercel (frontend)

Tech Stack

Backend

• Python 3.11
• Flask - REST API framework
• scikit-learn - Machine learning (Random Forest)
• yfinance - Real-time stock data
• pandas/numpy - Data processing
• gunicorn - Production server

Frontend

• React 19 - UI framework
• Tailwind CSS - Styling
• Recharts - Data visualization
• Axios - API client
• Lucide React - Icons

Project Structure

stocky/
├── backend/
│   ├── app.py                    # Flask API server
│   ├── train_model.py            # Model training script
│   ├── predict.py                # Prediction logic
│   ├── feature_engineering.py    # Technical indicators (30 features)
│   ├── data_fetcher.py           # Stock data downloader
│   ├── models/
│   │   └── spy_model.pkl         # Trained model
│   ├── requirements.txt          # Python dependencies
│   ├── Procfile                  # Railway deployment config
│   └── runtime.txt               # Python version
├── frontend/
│   ├── src/
│   │   ├── App.js                # Main React component
│   │   ├── index.js              # React entry point
│   │   └── index.css             # Tailwind styles
│   ├── public/
│   ├── package.json              # Node dependencies
│   ├── vercel.json               # Vercel deployment config
│   └── .env.example              # Environment variable template
├── DEPLOYMENT.md                 # Comprehensive deployment guide
├── RAILWAY_DEPLOY.md             # Quick Railway guide
└── VERCEL_DEPLOY.md              # Quick Vercel guide

Quick Start

Prerequisites

• Python 3.11+
• Node.js 16+
• Git

1. Clone the Repository

git clone https://github.com/YOUR_USERNAME/stock-predictor.git
cd stock-predictor

2. Setup Backend

cd backend

# Create virtual environment
python -m venv venv

# Activate virtual environment
# Windows:
venv\Scripts\activate
# Mac/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Train the model (optional - model is already included)
python train_model.py

# Run the server
python app.py

Backend will run at http://localhost:5000

3. Setup Frontend

cd frontend

# Install dependencies
npm install

# Start development server
npm start

Frontend will open at http://localhost:3000

API Endpoints

Health & Info

• GET /api/health - Health check
• GET /api/model/info - Model information
• GET /api/assets - Supported assets

Stock Data

• GET /api/predict/<symbol> - Get prediction for symbol
• GET /api/historical/<symbol>?period=1y&interval=1d - Historical data
• GET /api/info/<symbol> - Stock information
• POST /api/predict/batch - Batch predictions

Example Request

curl http://localhost:5000/api/predict/SPY

Example Response

{
  "success": true,
  "symbol": "SPY",
  "prediction": "UP",
  "confidence": 52.3,
  "current_price": 450.25,
  "probabilities": {
    "up": 52.3,
    "down": 47.7
  },
  "timestamp": "2025-10-09T..."
}

Machine Learning Model

Features (30 total)

• Price-based: Returns, gaps, streaks
• Moving Averages: SMA_5, SMA_20, SMA_50, EMA_12, EMA_26
• Volatility: ATR, Bollinger Bands position
• Momentum: RSI, MACD, Stochastic, Williams %R, ROC
• Volume: OBV, MFI (Money Flow Index)
• Trend: CCI, DI_Diff (directional indicators)

Performance

• Test Accuracy: 51.88%
• Cross-Validation: 46.9% (5-fold)
• Model: Random Forest (300 estimators, balanced weights)
• Training Data: SPY (2,708 days)

Why 52% Accuracy?

Stock prediction is extremely difficult. Even 52% is better than random (50%) and can be profitable with proper risk management. This is for educational purposes only!

Deployment

Deploy to Production

See detailed guides:

• DEPLOYMENT.md - Complete deployment guide
• RAILWAY_DEPLOY.md - Backend deployment (Railway)
• VERCEL_DEPLOY.md - Frontend deployment (Vercel)

Quick Deployment Steps

1. Backend (Railway)

   • Push to GitHub
   • Connect Railway to your repo
   • Set root directory: backend
   • Generate domain
   • Get backend URL

2. Frontend (Vercel)

   • Connect Vercel to your repo
   • Set root directory: frontend
   • Add env var: REACT_APP_API_URL=https://your-backend.railway.app/api
   • Deploy
   • Get frontend URL

Both platforms offer free tiers and auto-deploy on Git push!

Environment Variables

Backend (Railway)

• PORT - Auto-set by Railway
• FLASK_ENV=production - Optional

Frontend (Vercel)

• REACT_APP_API_URL - Backend API URL (required)

Example:

REACT_APP_API_URL=https://stock-api.up.railway.app/api

Testing

Backend Tests

cd backend
python test_api.py

Manual API Testing

# Health check
curl http://localhost:5000/api/health

# Get prediction
curl http://localhost:5000/api/predict/SPY

# Historical data
curl http://localhost:5000/api/historical/AAPL?period=1mo

Screenshots

Main Interface

• Stock symbol search
• Prediction display (UP/DOWN with confidence)
• Current price and probabilities
• Stock information

Historical Chart

• 3-month price history
• Interactive line chart
• Clean, modern design

Supported Symbols

The model is trained on SPY but can predict any stock:

• ETFs: SPY, VOO, QQQ, IWF, SCHD
• Stocks: AAPL, MSFT, GOOGL, TSLA, etc.
• Commodities: GLD, SLV
• Crypto: BTC-USD, ETH-USD

⚠️ Note: Model accuracy varies by symbol. Best performance on SPY.

Limitations & Disclaimer

⚠️ IMPORTANT: This is for educational purposes only!

• Not financial advice - Do not invest real money based on these predictions
• 52% accuracy - Only slightly better than random chance
• Past performance ≠ future results - Markets are unpredictable
• No guarantee - Model can be wrong, especially during volatile periods
• Use at your own risk - Always do your own research

Improving the Model

Want better predictions? Try:

1. More data: Train on 10+ years instead of ~5 years
2. More features: Add sentiment analysis, economic indicators
3. Better algorithms: Try LSTM, XGBoost, ensemble methods
4. Feature selection: Remove low-importance features
5. Hyperparameter tuning: Use GridSearchCV
6. More assets: Train separate models per stock
7. Different targets: Predict price ranges instead of binary UP/DOWN

Contributing

Contributions welcome! Feel free to:

• Report bugs
• Suggest features
• Improve the model
• Add new technical indicators
• Enhance the UI

License

MIT License - feel free to use for learning and personal projects!

Acknowledgments

• yfinance - Free stock data
• scikit-learn - Machine learning library
• Create React App - React setup
• Tailwind CSS - Styling framework
• Recharts - Charting library

Support

For issues or questions:

• Check DEPLOYMENT.md for deployment help
• Review API documentation in backend/API_GUIDE.md
• Check browser console for frontend errors
• Review Railway/Vercel logs for backend errors

What's Next?

After deployment:

1. ✅ Share your app with friends
2. ✅ Add more features (watchlists, alerts)
3. ✅ Improve the model accuracy
4. ✅ Add user authentication
5. ✅ Set up custom domain
6. ✅ Add more technical indicators
7. ✅ Train on more stocks

---

Built with ❤️ using Python, React, and Machine Learning

Happy Predicting! 📈🚀