Free, verified 1-minute financial data for backtesting and ML.
"Data Freedom for All" — In the age of AI, data must be open.
Alexandria is a free, open-source data infrastructure that provides:
- 📊 Verified 1-minute OHLCV data for BTC, NASDAQ, Dow Jones
- 🔥 99.95% correlation with Yahoo Finance (BTC)
- 💾 50:1 compression using ZSTD + Hive partitioning
- 🚀 M3/ARM optimized with Polars zero-copy reads
- 🩹 Brownian Bridge gap repair for missing candles
pip install -r requirements.txtimport polars as pl
# Read BTC sample data
btc = pl.read_parquet("data/BTCUSDT-1m-2024.parquet")
print(f"Loaded {btc.height:,} 1-minute candles")
print(btc.head())All data independently verified against Yahoo Finance:
| Symbol | Correlation | Direction Match | Source |
|---|---|---|---|
| BTCUSDT | 99.95% | 98.6% | Binance |
| NSXUSD | 90.14% | 87.8% | HistData |
| USA30IDXUSD | 91.00% | 88.9% | Dukascopy |
from alexandria import SniperDownloader
# Download BTC data (uses Sniper Mode - bandwidth efficient)
sniper = SniperDownloader()
offset, size = sniper.fetch_zip_metadata(url, "BTCUSDT-1m-2024-01.csv")
# See examples/02_download_data.py for full example| Module | Purpose |
|---|---|
core.py |
FoundryCore - Hive-partitioned storage |
harvester.py |
Binance Sniper acquisition |
harvester_tradfi.py |
HistData.com for indices |
harvester_dukascopy.py |
Dukascopy Bank tick data |
janitor.py |
Brownian Bridge gap repair |
features.py |
JIT indicator calculation |
Alexandria uses Brownian Bridge SDE (not linear interpolation) to generate statistically valid synthetic candles that preserve market volatility.
MIT License - Free to use, modify, and distribute.
Built with ❤️ for the open-source community.