markets.parquet
Market metadata, tokens, interval, price-to-beat metadata and capture state.
BTC 5-minute research data
Spend less time cleaning market data. Start backtesting.
BTC5m Data turns raw BTC 5-minute prediction-market captures into aligned, documented datasets with Full-L2 snapshots, quotes, reference prices, manifests and quality reports.
Independent research-data project. Not affiliated with or endorsed by Polymarket, Chainlink or any exchange. Public sample data is synthetic.
Forward collection running
- Market
- BTC 5m
- Book
- Full-L2 snapshots
- Timing
- UTC, source + receive timestamps
- Gate
- FULL_WINDOW only
- Output
- Parquet + JSON reports
What is inside
One 5-minute window, ready for analysis.
book_snapshots.parquet
Timestamped Full-L2 state snapshots for both outcome tokens.
book_levels.parquet
Flattened bid and ask levels tied back to each snapshot.
quotes.parquet
Best bid/ask quote series for easier event studies and execution research.
reference_prices.parquet
Aligned reference-price ticks with explicit provenance and redistribution status.
manifest + quality report
Checksums, row counts, coverage metrics, gap information and provenance.
Polymarket BTC 5m historical data
What a backtest-ready BTC 5-minute package needs beyond a final price.
Polymarket historical order-book data
When Full-L2 history matters, and how snapshots, levels and quality evidence fit together.
Quality first
A file existing does not mean it is usable.
Captures are admitted to inventory only when the recorder marks the window FULL_WINDOW and its coverage check passes. Partial captures remain visible internally as rejected evidence; they are not silently packaged.
✓SHA256 from raw capture to package manifest
✓Missing-rate and connection-gap checks
✓Row-count and Parquet integrity validation
✓No silent overwrite for conflicting source captures
✓Native / derived / missing provenance kept explicit
Early-access pricing target
Cheap enough to test before you commit.
For checking schema and research fit.
For a small backtest or pipeline trial.
For a more useful research sample.
Pricing is a demand-validation target, not an active checkout offer. Availability of real third-party-derived fields is subject to applicable redistribution rights.
Need a specific period?
Tell us the date range and fields you actually need.
Requests help decide what to collect, backfill and package next.
FAQ
What this is — and what it is not.
Is this a trading signal service?
No. It is historical/research data intended for analysis and backtesting.
Is the sample real market data?
No. The public sample is deliberately synthetic. Real-data publication and sale remain subject to licensing and redistribution permissions.
Do you guarantee every 5-minute window?
No. A failed or partial capture is rejected instead of being presented as complete.
Will you add ETH, SOL or an API?
Only if BTC 5m demand is validated first.