[Talk-us] Satus CDP

OSM Volunteer stevea steveaOSM at softworkers.com
Thu Mar 1 18:53:27 UTC 2018


Albert Pundt <roadsguy99 at gmail.com> writes
> Many towns and suburbs in my area are only CDPs, and having proper boundaries for them seems like it'd be useful, especially in more densely populated  areas. It's not like there's any fuzziness with them either, since they're defined by the Census Bureau and could only change once every 10 years. Only one U.S. census has occurred in OSM history, so it's not like we'd be constantly updating them. 

Thank you for your perspective Albert, and while you didn't ask a direct question, I am left with a couple myself after reading your observations:

Are these actually "towns" (or "cities") and so should be mapped boundary=administrative and admin_level=8?
Are these actually "suburbs" and so should be mapped as nodes tagged place=suburb inside of cities (which are mapped as the previous sentence)?

Noting in our United States admin_level wiki that a particular state SHOULD map with a particular "rungs on the ladder" hierarchy of particular admin_level values is useful, since by consensus we took our time to get those correct for that particular state.  THEN, there are assigning proper values on those proper entities, whether they came from the Census Bureau or need to be created/come from some other source.  If a census boundary exactly matches a city or town boundary, for example, (though it might prove challenging to discover or verify that), well, by all means:  rather than deleting that tagged polygon, we can simply change the tags from boundary=census to boundary=administrative, add an admin_level=8 tag (perhaps add a border_type=city tag) and be done until the next decennial census.

However, that's the tricky part: IS that Census Bureau-produced boundary truly the town or city boundary?  Or is it simply (and likely incorrectly) a "Census Bureau-produced boundary of census tract agglomerations" rather than a true corporate boundary as denoted by the city itself (or its parent state)?  Either might be correct, but as of now, data in our map are not sure.  In OSM, I'd like the data surely to be what it claims to be.  I don't believe we have that today with many Census Bureau boundaries, except that they denote a "boundary" of some sort:  sometimes correctly denoted "census" with no admin_level value (but should have one on a different and correct polygon), sometimes incorrectly denoted "administrative" with a questionable (maybe correct, maybe not) value, but on a polygon which is Census Bureau produced and possibly incorrect, possibly correct but we don't know that the Census Bureau exactly mapped a corporate boundary 

Locally speaking, they might appear to be "better than nothing, at least until the next decennial census," but ask yourself:  what are these really denoting?  If the answer is "we don't know if this is a corporate boundary or not" then we must at a minimum change the boundary tag from administrative to census, deleting any admin_level tag.  (Taking note amongst ourselves that these boundaries are marginally useful at best, especially when there are entities like towns and cities that deserve to have their corporate boundaries in our map).

We've already achieved consensus that "CDPs are lesser entities."  I'm suggesting we go ahead and delete them as noise (except in truly useful circumstances absent any "better" data, as in Alaska).  Superseding them with better data:  I'm all for that where we can do so.

SteveA
California


More information about the Talk-us mailing list