

Current coverage: 2025-07-01 to 2026-07-28
1.42 billion BTC/USDT trades from the most recent 12 months. Tick-level data in Apache Parquet, audited row by row. Delivered as a single ZIP containing all 4 quarterly Parquet files plus documentation and the Strategy Starter Kit notebook. Updated every 3 months - always the latest rolling year.
Includes a free Strategy Starter Kit (Jupyter notebook): VWAP, realized volatility, and order flow imbalance - pre-built, ready to run. Order flow imbalance can only be calculated with tick-level data, not OHLCV.
Schema (9 columns): exchange, symbol, trade_id, exchange_ts, ingest_ts, price, quantity, aggressor_side, raw_side
Format: Apache Parquet (ZSTD), 4 quarterly files. Total package 3.59 GB; single quarter 679 MB-1.3 GB.
Quality: Zero look-ahead bias, audited row counts, README and Data Quality Disclosure included. Per-file SHA256 checksum in MANIFEST.json.
Use cases: Backtesting, HFT simulation, market microstructure research, AI model training.
More at theglitchlist.com