Setup

1. Create virtualenv: python -m venv venv && source venv/bin/activate
2. Install: pip install -r requirements.txt

Files

• data_fetch.py # download historical OHLCV
• stream_data.py # (optional) live candle streamer
• features.py # feature engineering
• train_model.py # train and save LightGBM model
• backtest.py # quick backtest
• dashboard.py # Streamlit UI

Typical workflow

1. python data_fetch.py --symbol BTC/USDT --since_days 7
2. python train_model.py --symbol BTC/USDT
3. python backtest.py --symbol BTC/USDT
4. streamlit run dashboard.py

Note

• Uses Binance REST for historical bars. For production/low-latency, use websocket orderbook streams and Parquet/Arrow storage.