US Insider Trading Dataset 2006–2026: Every Form 4 & 5 Transaction, Cleaned

Every insider trade reported on SEC Forms 4 and 5 from January 2006 to October 9, 2026, cleaned and ready to analyze: 7.36 million transactions in one instant download, with quality flags on every row and a ready-made table of insider buying clusters. CSV and Parquet.

What's inside

• 7,360,555 transaction rows from 3,316,505 filings (2006-01-03 to 2026-10-09)

• 867,910 open-market purchases (transaction code P)

• purchase_clusters: 266,910 company-days with insider buying, including 40,830 days where two or more insiders bought

• One file per year, in CSV (gzip) and Parquet

• DATA_DICTIONARY.md, QUALITY_REPORT.json and SHA256SUMS checksums

What we fixed in the raw government files

• Full price precision for recent filings, parsed from the original XML (the raw quarterly files round prices to 2 decimals)

• Tickers cleaned ("NONE", "N/A" and junk removed), plus each company's current ticker

• CEO and CFO flags derived from officer titles• Joint filings resolved to one primary owner, with all owners listed

• Quality flags on every row: missing price, bad dates, price outliers, implausible values. 97.6% of rows pass every check, and nothing is silently deleted.

Built for

• Quant researchers backtesting insider-buying signals• Fintech apps and dashboards

• Academic studies, theses and journalism

Free sample

January 2024 transactions plus all 2024 purchase clusters: huggingface.co/datasets/Wetinkdata/us-insider-trading-form4-sample

License

One purchase covers you or your company. Use the data for anything: research, trading models, analysis, and inside your own products and services. Publish charts and findings freely. The only rule: don't resell, share or give away the dataset itself.The underlying facts are public-domain US government records; your purchase covers our cleaned, enriched compilation and its documentation.

Format

One zip download (564 MB). Data through October 9, 2026. Updated editions are released periodically.

Questions?

wetinkdata@gmail.com

Not affiliated with or endorsed by the SEC or any government agency. Not investment advice.