I keep four small datasets that I measure or maintain myself, and I publish them as CSV and JSON because people keep asking for the raw numbers behind the tables.
- e-mail security (SPF, DMARC, MTA-STS) — 102 rows, measured 2026-11-09: CSV · JSON
- HTTP security headers (HSTS, CSP, COOP) — 98 rows, measured 2026-11-09: CSV · JSON
- version support and end-of-life dates — 627 rows, measured 2026-11-11: CSV · JSON
- error messages with cause and fix — 149 rows, measured 2026-11-10: CSV · JSON
Every file carries its own scan date, the column definitions and the licence (CC BY 4.0). No signup, no rate limit, no tracking — a plain file behind a plain URL.
Two of them are measured from the outside: public DNS for SPF/DMARC/MTA-STS, the response headers of the homepage for HSTS/CSP/COOP. The other two are curated: end-of-support dates per version cycle and 149 error messages with cause, checks and fix.
If you cite them, the suggested citation is in the hub. If you find a wrong reading, the check is reproducible — everything is a single DNS query or a single curl -I.
Hub with column definitions, scan dates and citation: daniele-michelin.com/en/open-data/
First published at daniele-michelin.com. Four datasets, measured weekly, CC BY 4.0 — CC BY means the numbers are yours to reuse, the link is the only price.
