[Talk-us] Strange city boundary: Lee, Illinois

OSM Volunteer stevea steveaOSM at softworkers.com
Wed Nov 14 20:02:01 UTC 2018

Carl Anderson is correct:  what is in the map from TIGER about LSAD is true and affords the possibly to derive geo data about incorporated entities (in some cases, where they haven't been deleted), although the data (being somewhere between 11 and 13 years old) may not be accurate, given annexations, etc.  However, OSM's community, through exhaustive consensus (much of it right here on talk-us, many of these discussions are ref'd in a wiki I noted earlier) agree that what the US Census Bureau says is not necessarily what OSM does or should use to document such entities.  In other words, the Census Bureau is not authoritative.  For what we agree to put into OSM, the OSM community's consensus IS authoritative.  We have agreed that census boundaries of CDPs are statistical, not administrative (what admin_level attempts to denote).  We correctly document how to do this.  However, MUCH old TIGER data remain.

Martijn, your OT link is helpful, here is a visual version:  http://overpass-turbo.eu/s/DG5 (although that does not allow "census.gov" to appear as often as your text-based version does, so thank you for that formatting).  What this shows is that the importation of much Census Bureau data as CDPs in Utah (and elsewhere in the USA) continues to have MUCH work to do:  our wiki suggests admin_level=8 tag be removed from these, and the boundary=administrative tag be changed to boundary=census.  This is correct, it has achieved wide consensus in OSM (via these talk-us pages) and has been documented in our wiki for some time.  And not simply in Utah, this is true in all 50 states, except Alaska, where the state works closely with the Census Bureau to "organize" (not in the legal sense) the Unorganized Borough of Alaska (bigger than any other state, even Texas).  Carl offers a clever way for us to sharpen up how we might do this:  choose admin_level=8 tagged relations which have tiger:LSAD=57 (e.g. Mona, Utah), change boundary=administrative to boundary=census, and delete the admin_level=8 tag.  Import script, anyone?  (Whew, that's a loaded question!)

However, I disagree with Martijn (or perhaps I do not understand his intention) as he says about our US_admin_level wiki "United_States_admin_level is really not correct... CDPs should be tagged boundary=census, ideally without an admin_level=* tag."  I believe that IS a correct statement, it is simply that Utah (and many other states, again, not Alaska) have never had this "clean up" done.

Martijn's assertion that "state municipalities: cities, towns, villages and hamlets (infrequent)” is an incorrect description of what we INTEND to denote with admin_level=8 in the USA is also incorrect.  Simply said, hoary old TIGER data contradicts this true statement on a fairly large scale, in Utah, yes, but in most other states, too.

Let's not confuse what's in the map (from a noisy TIGER import) as correct, when what really is correct is what we have achieved consensus about and wiki-documented:  CDP data are boundary=census, not boundary=administrative (and so, should have NO admin_level key, with any value).

And, I'd much rather have "too much" wiki than "not enough."  Wiki can be ignored if too verbose, however, the consensus such wiki express is not easily conjured.

Where I agree with Martijn is "I guess we have some work to do."  "Clean the catbox" indeed!


More information about the Talk-us mailing list