v0.2.0 – Enhanced Microstructure Analysis

Release v0.2.0

🚀 Highlights

• Enhanced distribution fitting

  • Spread: lognormal, gamma, exponential, Pareto
  • Volume: lognormal, gamma, Pareto
  • Returns: Student-t, Laplace, Normal
  • |Returns|: exponential, lognormal, gamma

• New visualizations

  • QQ plots (lognormal, Student-t, etc.)
  • Log-log tail plots for heavy-tail inspection
  • Intraday heatmaps (spread or |returns|)
  • |Returns| autocorrelation for volatility clustering

• HTML report improvements

  • Cleaner table formatting with rounded parameters
  • Grid layout improved (3 figures per row)
  • Key notes section summarizing findings

• Data robustness

  • Auto-detect headerless Binance aggTrades CSV (8-column format)
  • Support for both CSV and Parquet as input
  • Cleaning works with only trades, only book, or both

• Engineering updates

  • Windows-compatible entrypoint (python scripts/run_all.py)
  • Simulated book generator (convert_trades_to_book.py) for trades-only users
  • Sample data pipeline fully supported with --use-sample

📂 Assets

• Figures → results/figures/
• Tables → results/tables/*.csv
• Report → reports/summary.html