📈 Stock Market Analyzer

A production-quality Python project for fetching, analysing, and visualising stock market data — complete with technical indicators, ML-based price prediction, a backtesting engine, and an interactive Plotly/Dash dashboard.

---

🗂 Project Structure

stock_analyzer/
├── app.py                  # Dash dashboard entry point
├── requirements.txt
├── pytest.ini
├── src/
│   ├── __init__.py
│   ├── data_fetcher.py     # yfinance wrapper with caching
│   ├── indicators.py       # Technical indicators + signal generator
│   ├── backtester.py       # Event-driven backtesting engine
│   ├── predictor.py        # Random Forest / GBM price direction model
│   └── charts.py           # Plotly chart builders
├── tests/
│   ├── test_indicators.py
│   └── test_backtester.py
└── .vscode/
    ├── settings.json
    └── launch.json

---

⚡ Quick Start

1. Clone / open in VS Code

cd stock_analyzer
code .

2. Create virtual environment

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

3. Install dependencies

pip install -r requirements.txt

4. Run the dashboard

python app.py

Open http://localhost:8050 in your browser.

5. Run tests

pytest tests/ -v

---

🔧 Features

Module	What it does
data_fetcher.py	Downloads OHLCV + metadata via yfinance; in-memory cache; multi-ticker support
indicators.py	SMA, EMA, RSI, MACD, Bollinger Bands, Stochastic, ATR, OBV; composite signal generator
backtester.py	Signal-driven backtest; tracks trades, equity curve, Sharpe, max drawdown, win rate
predictor.py	RandomForest / GradientBoosting next-day direction prediction; time-series CV
charts.py	Candlestick + volume, RSI, MACD, equity curve, multi-stock comparison (all dark-themed Plotly)
app.py	Dash dashboard wiring everything together with live KPI cards and comparison mode

---

🧩 Extending the Project

Add a new indicator

# src/indicators.py
@staticmethod
def add_vwap(df: pd.DataFrame) -> pd.DataFrame:
    df = df.copy()
    typical_price = (df["High"] + df["Low"] + df["Close"]) / 3
    df["VWAP"] = (typical_price * df["Volume"]).cumsum() / df["Volume"].cumsum()
    return df

Add a new strategy

# Create a custom signal column before calling Backtester.run()
df["MySignal"] = 0
df.loc[df["RSI"] < 25, "MySignal"] = 1   # Very oversold → BUY
df.loc[df["RSI"] > 75, "MySignal"] = -1  # Very overbought → SELL

bt = Backtester(initial_capital=10_000)
result = bt.run(df, signal_col="MySignal")
print(bt.summary(result))

Switch ML model

predictor = PricePredictor(model_type="gradient_boost")
predictor.train(df)
direction, confidence = predictor.predict_next(df)

---

📚 Key Libraries

Library	Purpose
yfinance	Stock data from Yahoo Finance
pandas / numpy	Data manipulation
ta	40+ technical indicators
scikit-learn	ML models + preprocessing
plotly	Interactive charts
dash	Web dashboard framework
dash-bootstrap-components	UI components

---

🧪 Testing

Tests use pytest with synthetic data (no network calls required):

pytest tests/ -v               # All tests
pytest tests/test_indicators.py  # Just indicators
pytest -k "TestRSI"            # Filter by class/name

---

🚀 Ideas for Further Development

• Add portfolio optimisation (Markowitz / Sharpe maximisation)
• Integrate live WebSocket price feeds
• Add SQLite persistence for historical backtest results
• Deploy to Render / Railway as a web app
• Add email/Telegram alerts for signal changes
• Implement LSTM / Transformer for time-series prediction
• Add options chain analysis