[Talk-us] deleting misleading CDPs

stevea steveaOSM at softworkers.com
Fri May 31 18:50:36 UTC 2013


I agree with your particular case that this CDP (Census Designated 
Place) might be deleted, mostly for the reason that there is a named 
town with the same name.  A named town has specific borders codified 
in state or local statute, making it distinctly real and distinctly 
of local importance.  Given the choice to map town or CDP, I'd say 
the town is the better choice among the two.

A CDP, on the other hand, is something which is federal in nature, 
and which might be deprecated properly in light of the 10th amendment 
giving states (and localities like counties, which are divisions of 
states) rights where the Constitution does not delineate federal 
powers.  But the Constitution DOES delineate the need for a census, 
so CDPs really might (in a legal, constitutional sense) exist for a 
"good reason."  We have a certain state-and-federal system here in 
the USA, but we do have it.

In the instant case, I'm (barely) OK with the deletion, as it causes 
confusion the way it is now.  HOWEVER, a better solution may be to 
draw the town, name it, and KEEP the CDP, renaming it "Niskayuna CDP" 
making it clear that it is a federally-designated area not exactly 
the same as the town.  Which is true.  (Coding for the renderer? 
Yes, this leans that way, but in the interests of clarity, so I'm OK 
with it).

CDPs were discussed in 
http://wiki.openstreetmap.org/wiki/United_States_admin_level and 
Talk-us Digest, Volume 62, Issue 1 ("An admin_level for CDPs?").  A 
consensus that seemed to emerge was "CDP polygons imported from TIGER 
data should not be tagged boundary=administrative (implying an 
admin_level tag) but should rather be tagged boundary=census. 
Accordingly, no admin_level tag is required on CDP polygons. (Not to 
mention a lot of work to update them, whether manually or by script)."

In Ohio CDP boundaries are being retagged with boundary=census and 
place=locality but without admin_level.  Hence, they still show up in 
Nominatim as localities:  both useful and correct.

Importantly, Minh Nguyen writes:  "I'm not fundamentally opposed to 
putting in statistical areas; I just think it may be less confusing 
to use some other value of boundary=* (even with admin_level set), 
rather than overloading boundary=administrative for what evidently 
isn't a straightforward hierarchy of government entities. It's 
specialized information, less important than your typical city/county 
distinctions when completing the sentence "This business is located 
in..."  To which I agree.

Then I said:  "What I found useful to do around here (where there are 
CDP polygons entered from TIGER, but they have no admin_level tag) is 
to add a point tagged hamlet=* or village=* or town=* (but not 
necessarily suburb=* as that implies city subordination, nor city=* 
as that implies incorporation) to the "approximate center point" of 
the CDP polygon, along with a name=* tag that matches the name of the 
CDP. This point might logically be a mathematical centroid, but I 
have found it more useful to place this point at a more culturally 
significant point in the "human center" of the community designated 
by the CDP.  Usually this is at or near a significant crossroads, 
where there might be a market, a church, a school, a small commercial 
district, or the like."

To which Minh replies:  "Yes, this makes a lot of sense. TIGER 2008 
came with place=hamlets for all the 2010 CDPs in the Cincinnati area, 
all in very sensible locations, so I just assumed the CDPs were a 
subset of all the unincorporated areas in TIGER."  Of course, again, 
I would agree with keeping the place= POIs.

Given all this, what I would do is rename the CDP "Niskayuna CDP," 
delete its admin_level tag, add the tag boundary=census, AND draw the 
town boundaries with an admin_level=8 and name=Niskayuna.  If you 
want to additionally "force a render" of a cultural centerpoint of 
Niskayuna CDP (to be sure:  distinct from the town), you could add a 
node with name=Niskayuna CDP and place=[hamlet, village, town] as 

I hope this all helps.


>i'm planning to delete a misleading CDP in the near future, i'm pondering
>the fact that from time to time we talk about deleting all the CDPs, an idea
>which i sometimes think is the right idea.
>in this case, the CDP is for Niskayuna NY
>which is a tiny sliver of the Town of Niskayuna. the CDP is bounded 
>on the east
>by Balltown Road, on the west by the level 8 admin boundary about 8 
>blocks away,
>on the south by Union Street, and on the north by Providence Avenue.
>at the same time i will add the proper border for the Town which is not there
>at present. nobody even knows what the CDP is and it's mostly just confusing.
>the town offices and the high school aren't even inside the CDP boundary.
>thoughts, anyone?
>my feeling is that if there's a named town then including a much smaller CDP
>with the same name is quite misleading. i think the same situation exists in
>Rotterdam NY and if i find that that's the case, i'll apply the same remedy.

More information about the Talk-us mailing list