[Talk-us] Pittsburgh neighborhood boundaries mapped with admin level 9?

OSM Volunteer stevea steveaOSM at softworkers.com
Thu Jul 27 17:22:49 UTC 2017

> Albert Pundt <roadsguy99 at gmail.com> writes
> I noticed that the neighborhoods in Pittsburgh are mapped as administrative boundaries with admin_level=9. Is this proper? The wiki page for U.S. admin levels doesn't list any use for admin level 9 in Pennsylvania, though this seems appropriate if Pittsburgh neighborhoods are true administrative divisions. It just needs to be documented, or perhaps used elsewhere in the state, like with the fairly distinct neighborhoods in Philadelphia.

Albert, I noticed this, too.  There has been recent development of USA admin_level:  the wiki page you note has been undergoing harmonization, with the result that as of about three weeks ago, "about 41 out of 50" states (leaving 9) were believed correct as described in that table.  (And, that was posted here).  On July 10, Peter Dobratz and I (among others) submitted the six New England states (of those 9) in a sub-table, which I harmonized into that wiki's "Big Table."  (Pennsylvania was not included then).  That brought us to 47 states having their values in the Big Table believed correct, with Illinois, Minnesota and Pennsylvania being the three remaining.

The next day, Illinois was aligned with its Wikipedia article describing "precincts" in the west central and southern part of the state (so "Precinct" was entered in Illinois' row in the Big Table; thanks, MikeN).  Minnesota now has a footnote that while the table is believed correct (Townships exist at admin_level=7), our map does not contain any (or many) entered Minnesota Township boundaries.  (But it could, should intrepid mappers wish to enter them).  That leaves Pennsylvania as the sole state of 50 needing to go through this "OSM vetting/review of whether its Big Table row is correct," and here we are.  At least for Pennsylvania's city's neighborhoods.

Albert, one possibility is that you "assert" that admin_level=9 tags on Pittsburgh neighborhoods are correct, decide to document this in our wiki by modifying the Big Table to say so (wiki-editing these tables can be tricky) and leave the neighborhoods tagged as they are now in the map.  So when you ask "is this proper?" I'm not sure if you are asking if 1) these neighborhoods actually exist as "government units," 2) whether 9 is the right value or maybe 10 is correct, or 3) whether it only SEEMS improper because the Big Table doesn't list "Neighborhood" in Pennsylvania's row in the 9 column.  If these are real neighborhoods, and tagging them 9 is what is going on "on the ground," then I'm OK with "backfilling" the table with that entry, as it asserts what the rest of the Big Table now does:  "Here is what OSM believes admin_level tagging SHOULD BE (not necessarily is entered into the map) on all fifty states, DC, five territories, and all their subdivisions."  You can change that table entry, or if you ask me to do it because you believe it correct (it is Pittsburgh's on-the-ground reality), I can do it.  Not just Pittsburgh, but Philadelphia having neighborhoods, either in real life or tagged in the map, is also credible evidence to enter the value "Neighborhood" in Pennsylvania's Big Table row entry for the 9 value, too.

However, OSM has been selecting 10 instead of 9 for neighborhoods, leaving something like 9 for Precinct or Ward, this is true in New Orleans (but unentered in OSM), which has both, for example.  So while I am simply "one (informed) observer" here, (with a concomitant "only one vote" for this), I'd suggest that Pittsburgh neighborhoods be retagged with admin_level=10 instead of 9.  Again, that's just me, but there is good precedent to do so.

It should be reiterated that US_admin_level wiki Big Table entries are "believed correct values" (through research, a Wikipedia explanation, law, OSM consensus, or best:  some good combination of all of these) AND that there are places in our map where the myriad of subdivisions listed there (e.g. townships in Minnesota) have not been entered into OSM, but could/should be.  The values of 2 (federal/national) and 4 (state) are believed 100% correct, as are perhaps 99+% of value 6 for county or county equivalent.  Many values of 7 (township) are missing, that is a frontier of effort for OSM to continue to complete.  Most values set to 8 (city/town) are entered and correct, though we likely need more.  And boundaries with admin_level 9 and 10, representing wards, precincts and neighborhoods (where they truly exist as governmental units) are another frontier which need to be identified and entered, though OSM is doing that work right now (to wit).  Some minor discussion/tune-ups (like this) should not surprise or alarm us.

The good news is that the Big Table can be said to be believed by OSM consensus to be "very nearly correct."  I'd guess only very minor changes to it will be required going forward.

I endeavor to be crystal clear here, but these can be complicated topics to explain, so if you have further questions, please email me off-list.


More information about the Talk-us mailing list