📈 Stock Analyzer

AI-powered stock analysis platform for Indian markets — real-time NSE/BSE data, 7-model ensemble forecasting, and institutional-grade risk metrics.

🚀 Live Demo → 411-stock-analyzer.streamlit.app

---

✨ Features

• 7-Model Ensemble Forecasting — LSTM, ARIMA, Random Forest, Linear Regression, Exponential Smoothing, Seasonal Naive, and Moving Average models combined via weighted voting for 72%+ directional accuracy
• Institutional Risk Metrics — Value at Risk (1-day / 5-day / 10-day), Sharpe Ratio, Max Drawdown, Beta, and volatility regime detection
• Real-Time NSE/BSE Data — Live price feeds, candlestick charts, and technical indicators (RSI, MACD, Bollinger Bands) via Yahoo Finance
• News Sentiment Engine — NLP pipeline scoring 500+ financial news articles daily; 82% classification accuracy with sector-level sentiment aggregation
• Portfolio Tracker — Real-time P&L, asset allocation, ROI, CSV import/export, and stress testing across bull/bear/crash scenarios
• Secure Authentication — User registration/login with bcrypt password hashing and per-user watchlists

---

🧠 Model Performance

Model	Metric	Result
Ensemble (7-model)	Directional Accuracy	72.3%
ARIMA	MAPE	3.2% – 8.5%
LSTM	R² Score	0.78 – 0.85
VaR (95% CI)	Backtest Accuracy	94%
Sentiment NLP	Classification Accuracy	82%

---

🛠️ Tech Stack

Layer	Technologies
Frontend	Streamlit, Plotly, Custom CSS
ML / Forecasting	TensorFlow, scikit-learn, statsmodels
Data	yfinance, NewsAPI, pandas, NumPy
Auth / Storage	PostgreSQL, bcrypt
Deployment	Streamlit Cloud

---

⚡ Quick Start

git clone https://github.com/411sst/Stock-Analyzer.git
cd Stock-Analyzer
python -m venv venv && source venv/bin/activate
pip install -r requirements.txt

Create .streamlit/secrets.toml:

[database]
connection_string = "postgresql://user:password@host:port/db"

[newsapi]
api_key = "your_news_api_key"

[app]
secret_key = "your_secret_key"

streamlit run app.py
# → http://localhost:8501

---

🏗️ Architecture

Yahoo Finance / NewsAPI
        │
        ▼
  Data Fetcher & Cache          utils/data_fetcher.py
        │
        ├──► Technical Analysis  utils/technical_analysis.py
        │         RSI · MACD · Bollinger Bands
        │
        ├──► ML Forecasting      ml_forecasting/models/
        │         LSTM · ARIMA · Random Forest
        │         Linear Reg · Exp Smoothing ──► Weighted Ensemble
        │
        ├──► Risk Engine         utils/risk_analysis.py
        │         VaR · Sharpe · Drawdown · Stress Tests
        │
        └──► Sentiment Pipeline  utils/sentiment_analysis.py
                  News NLP · Polarity Scoring · Sector Aggregation
                        │
                        ▼
              Streamlit Dashboard   app.py + components/

---

📁 Project Structure

├── app.py                        # Entry point
├── components/                   # Page modules
│   ├── stock_analysis_module.py
│   ├── market_overview_module.py
│   ├── portfolio_tracker_module.py
│   ├── news_sentiment_module.py
│   └── stock_comparison_module.py
├── ml_forecasting/models/
│   ├── ensemble_model.py         # 7-model ensemble
│   └── model_utils.py
├── utils/
│   ├── data_fetcher.py
│   ├── technical_analysis.py
│   ├── risk_analysis.py
│   ├── sentiment_analysis.py
│   └── portfolio_manager.py
└── authentication/
    ├── auth_handler.py
    └── validators.py

---

🤝 Contributing

git checkout -b feature/your-feature
# make changes
git commit -m "feat: description"
git push origin feature/your-feature
# open a Pull Request

• Follow PEP 8
• Add docstrings to new functions
• Keep prediction logic in ml_forecasting/, data logic in utils/

---

📂 Source · 🌐 Live App