City polygon layer has updated (NA)

[paulwelch23]

The offending 'city' is 'M, Ohio'. If you scroll around the borders, especially with the help of the WME highlighting tool if you are an AM, you'll be able to identify and reassign those segments.

This update got rid of the cities of G and B, Misisssippi in the suburbs of Gulfport and Biloxi. I suspect we've all done this...probably more with city names that begin with letters that can be touch-typed with the left hand. I've never had it happen with more than one typed letter. (For New Orleans, I type "ew" and hit enter, Metairie, "et", Kenner "er", etc.")

We now have a multi-hundred, if not thousand square mile area of TN that is shown as "M, TN". This was created by (as best as I can tell) THREE segments. One in Murfreesboro, one in Mount Juliet, and one in Martin. This polygon covers the ENTIRE city of Nashville as well as many of of it major suburbs.

Great. Just great.

I got the segments in Murfreesboro and Mount Juliet fixed, but Martin in outside of my AM range. Can a CM fix this segment in Martin? Thanks in advance.

Maybe another update, Waze? Pretty please?

[paulwelch23]

Okay, I fixed that segment. Post here again when the "M" goes away. I've already fixed everything here and it's updated, so I'm curious to know if they're rebuilding the city polygons nightly or weekly or what.

Thanks, and I let you know. I'm very interested in this update schedule, too.

[paulwelch23]

Okay, I fixed that segment. Post here again when the "M" goes away. I've already fixed everything here and it's updated, so I'm curious to know if they're rebuilding the city polygons nightly or weekly or what.

Thanks, and I let you know. I'm very interested in this update schedule, too.

I can confirm that the city layer has been updated and that this is fixed. Woo hoo!

[scruffy151]

Here's a good one...

https://www.waze.com/editor/?zoom=4&lat ... TTTFTTTTFT

I noticed "D'Iberville" on the map twice, really close together, so I turned on the City Layer and there's this extra polygon. Using Color Highlights, D'Iberville is in there twice.

Took me a little while to figure it out, but the font on the map puts the little cross at the top and bottom of the capital i and that's when I saw the difference. Someone had named about a dozen segments "D'LBERVILLE" (caps here so you can read it).

Opps. I thought I sent you a PM about this. I cleaned up all the regular streets, but there a few left on I-10 and I-110 that I don't have permissions to change.

[scruffy151]

And D'Lberville is gone now...only about 10 hours after I removed it from the editor!

Yeah!!! One choice in the drop down menu.

[skbun]

Thought I'd throw this out there as an FYI to city and state smear checkers - looks like the city layer in the NA tiles has updated again, at last! First time in months.

(I'd personally like to throw my hat in, still for say 'City layer updating much more often and probably more often than the map tiles themselves run - to minimize pollution', or 'Not relying on a city's origin point but rather nearby segments only for valid city dropdowns ( http://www.waze.com/forum/viewtopic.php?f=10&t=31917 ) '

...but thank you, Waze. I hope this means you've been reading the city and state smear threads and are thinking about them.
http://www.waze.com/forum/viewtopic.php?f=129&t=23139 , http://www.waze.com/forum/viewtopic.php?f=10&t=32060 ,
http://www.waze.com/forum/viewtopic.php?f=10&t=31918 ...

[skbun]

So I threw on the city layer today and saw this:
This whole area is covered in a green city layer and it stretches from Cleveland down to Columbus & Cincinnati and over to Toledo. The city labels seem to be in the correct place but many city boundaries seem to have disappeared.

Hmmm. Okay. Found the border of the green city by scrolling around, and came to this permalink (undoubtedly one data point of many).

https://www.waze.com/editor/?zoom=8&lat ... FFTFTTTTFT
https://www.waze.com/editor/?zoom=8&lat ... FFTFTTTTFT
...and so on...

The offending 'city' is 'M, Ohio'. If you scroll around the borders, especially with the help of the WME highlighting tool if you are an AM, you'll be able to identify and reassign those segments.

Waze? I can't help but notice that both of these occurred in cities beginning with M, Marblehead, and Maumee. It's pointing to a problem I've had myself, with the WME city completion. Cities are given in the dropdown using cities.get, and one is highlighted. A user will often assume that hitting tab or similar will autocomplete it, but it won't. Instead, you get...this. I don't really think this is the user's fault. The editor can be clumsy about this.

Can we make the 'Cities' blank into an explicit 'Select' choices dropdown, whose last option is 'New city', where THEN a user gets a blank where they can fill in their own field? Or, if not that, at least just pop up a warning saying 'You are trying to create a new city, 'M, Ohio'. Are you sure you want to do that?'

[skbun]

This update got rid of the cities of G and B, Misisssippi in the suburbs of Gulfport and Biloxi. I suspect we've all done this...probably more with city names that begin with letters that can be touch-typed with the left hand. I've never had it happen with more than one typed letter. (For New Orleans, I type "ew" and hit enter, Metairie, "et", Kenner "er", etc.")

Oh, totally with you! The only reason I realized what's going on here is that I accidentally created Fe, Washington recently, meaning to complete 'Federal Way'. Heehee. The difference was, I caught it using WME highlighter city, and was 'Heeeey, why is this new city here. Oh. Oops,' and immediately fixed it.

It occurs to me too...this is another reason to run the city layer script 'much more often'. It's now apparent that 'segment too far from city' errors are not generated by nearest segment, but by polygon distance, and M, Ohio is a multipointed polygon now for that reason.

Something like...'If no polygon exists for this city, assume it's close enough to allow' is what it's doing now.

'If no polygon exists, and this segment is more than 20km/some-reasonable-distance from the origin segment where this city was created, just assume it's too far' would do it, IMHO. (Anyway, if the city layer runs much more often, the window for this problem to occur closes fast, and such a check is less necessary.)

[skbun]

Are you using the Colour Highlights script? If so, can the script detect and highlight that segment where the offending city is in the alternate field?

It should be able to. I wrote the portion of the code that does that for 1.5. It should appear as a dashed yellow line rather than solid, as Tim put it in. But it's also worth checking in the area you see such a city in the city dropdown list, to make sure you're also looking at 'the right city'.

I've seen cases where there are three different cities with the same name in a location - two with different states...heck, one with none at all.

[skbun]

Still appearing on the map are the neighborhood of "Blue Hills" (all segments removed over 6 months ago), and "Canaan" (the neightborhood in the town of North Canaan") which was removed (relabeled "North Canaan") so we could label the actual Town of Canaan as "Canaan" as it was never labeled in the first place.

Yeah, you can't get rid of those by removing all the segments. I have one here that's a CDP that nobody actually refers to, and I changed them all over a year ago. There's no more geometric outline, but the name still appears on the map. There's two city layers...one that we can have an effect on, and one that came from the basemap that's not dependent on segments with city names.

Another like this was brought to my attention as I did as second scan through the Seattle metro area, checking for blotches. Part of the 'basemap layer map' we can't change includes the CDP of Lea Hill, Washington, at https://www.waze.com/editor/?zoom=1&lat ... FFTFTTTTFT . It was a CDP until the end of 2007, at which time it was annexed by the city of Auburn. This can be seen on the interactive TIGER map for 2010. The reason I bring this up is not to bring up 'another example, etc. etc.', but to point out that the basemap sticky city layer is at least five years out of date, probably taken from the 2000 basemaps.

Certainly the 2010 basemaps would be available; there is also this:

http://www.census.gov/geo/partnerships/bas.html - the BAS, (Boundary and Annexation Survey), done annually, which includes shapefiles. If Waze does want to keep a locked city layer based on shapes, this would be a good place to go.