Been writing #openstreetmap USA phone number fixups for #josm for a bit. They’re thorough enough now that the residual tags often have some excellent surprises.
Check out this fun one:
The rules live in this repo if you want to go hunting for fun things. https://github.com/watmildon/josm-validator-rules
@watmildon My hobby is postal codes in OSM. And there you see japanese numbers (kanji keyboard) which have a different unicode symbol than ascii numbers. Also endash -, emdash — instead of – Or the occasional O vs 0. I encountered non-printable characters. I enjoy those typos. I cleaned up USA data in 2021, there’s always more one can do but all the easy patterns are done I think. Like removing state codes from addr:postcode ("CA 91210")
@mtmail Yes! Always more cleaning. I'm trying to keep things somewhat reproducible by someone just using JOSM so I've dumped some validator rules into a github repo. The phone number fixup is this one: https://github.com/watmildon/josm-validator-rules/blob/main/rules/PhoneNumber-ITU-T-Formatter.validator.mapcss