Skip to content

Tutorials

Interactive Jupyter notebooks demonstrating SignalFlow capabilities. All tutorials use synthetic data via VirtualDataProvider -- no API keys required.

Notebooks

01 - Quick Start

Build and run your first backtest in 5 minutes using the fluent builder API. Covers sf.Backtest(), sf.backtest(), and BacktestResult.

02 - Custom Detector

Create your own signal detector from scratch. Covers the SignalDetector base class, @sf.detector registration, multi-detector strategies, and signal aggregation.

03 - Data Loading & Resampling

Load market data from DuckDB stores, auto-detect timeframes, resample OHLCV data, and work with exchange-specific timeframe support.

04 - Pipeline Visualization

Visualize your backtest pipeline as an interactive D3.js graph or Mermaid diagram. Covers sf.viz.pipeline(), sf.viz.features(), and the local development server.

05 - Advanced Strategies

Build multi-detector ensemble strategies with signal aggregation, named entry/exit rules, position sizing, and YAML configuration.

Running Locally

# Install dependencies
pip install -e ".[dev]"

# Launch Jupyter
jupyter lab

All cell outputs (charts, tables, images) will be rendered automatically on the documentation site.