{"id":42579,"date":"2026-09-12T19:27:08","date_gmt":"2026-09-12T17:27:08","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/self-promotion-paid-i-parsed-the-full-nppes-provider-registry-9-27m-rows-into-plain-english-professions-free-1000-row-sample\/"},"modified":"2026-09-12T19:27:08","modified_gmt":"2026-09-12T17:27:08","slug":"self-promotion-paid-i-parsed-the-full-nppes-provider-registry-9-27m-rows-into-plain-english-professions-free-1000-row-sample","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/self-promotion-paid-i-parsed-the-full-nppes-provider-registry-9-27m-rows-into-plain-english-professions-free-1000-row-sample\/","title":{"rendered":"[self-promotion] [paid] I Parsed The Full NPPES Provider Registry (9.27M Rows) Into Plain-English Professions \u2014 Free 1,000-row Sample"},"content":{"rendered":"

<\/p>\n

\n

Every US healthcare provider has an NPI, and CMS publishes the whole registry as a bulk file. It’s public, but it ships as a multi-million-row monster with taxonomy codes instead of job titles, no category labels and no dedupe, so most people bounce off it.<\/p>\n

I mapped the taxonomy codes to 290 plain-English profession labels across 14 sectors, normalized the addresses and deduped it. 9,270,029 rows. Per row: name, full mailing address, profession, the federal source it came from, and the date it was last rebuilt.<\/p>\n

A few things I learned doing it:<\/p>\n