Quick Start¶
Get Quant101 running locally in 5 minutes.
Prerequisites¶
- Python 3.12+
- Poetry for dependency management
- Polygon.io flat file data (or any OHLCV source in Parquet)
Installation¶
Configuration¶
Copy the example config and edit your data paths:
Key fields in basic_config.yaml:
data:
data_dir: /path/to/your/polygon_data # Root of your data directory
update:
mode: standalone # standalone | server | client
Run the Test Suite¶
Run a Pipeline Backtest¶
The fastest way to see the system work end-to-end:
This runs the default alpha pipeline (BBIBOLL factor, equal-weight sizing, daily rebalancing) on the US Large Cap 50 universe and prints a Sharpe ratio.
Run a Strategy Backtest¶
For the legacy trade-level BBIBOLL backtest:
Explore Notebooks¶
The research notebooks are the best way to understand each module:
| Notebook | What It Shows |
|---|---|
pipeline_demo.ipynb |
Full pipeline: OHLCV → factor → weights → Sharpe |
factor_diagnostics.ipynb |
Per-factor IC/IR, L/S returns, direction check |
risk_analysis.ipynb |
VaR, CVaR, drawdown, 4 sizing methods compared |
cost_analysis.ipynb |
Transaction costs, breakeven analysis, rebalancing frequency |
validation.ipynb |
Walk-forward, bootstrap CI, PSR/DSR, multiple testing |
alpha_research.ipynb |
End-to-end BBIBOLL factor analysis |
alpha_iteration.ipynb |
Multi-factor: STR, Vol Ratio, IC correlation |
Next Steps¶
- Read the Architecture Overview to understand how modules connect
- Follow the Factor Research Workflow to build your own alpha
- Review the Lessons Learned to avoid common pitfalls