Hi James,
happy to help, and yes, I can see your changes just fine 
I have no knowledge of ruby on rails so apologies in advance if this is a stupid question, but I was wondering if it would make a difference in speed to make the ‘positive’ query, since there are more highway values to ignore than there are to keep.
Possibly!
I’ll have to figure out how to build that query. Maybe something that builds from:
['highway']
['highway' == 'primary']
This is all Overpass API stuff - the only Rails-specific code is the #{area_id + 3600000000} piece that generates the correct Overpass Area ID from the OSM Relation ID that I store locally.
I included a pile of links in the Overpass Query sheet in your doc.
I would consider adding [‘service’ !~ ‘drive-through’] to the list. I have several Dunkin’ Donuts, Burger King, or other food drive-thrus showing up as streets in CityStrides.
I’m surprised that your local Dunkin’ drive-through actually got a name= tag…
But yes, I quite agree. Re-reading the OSM docs, I might even go one step further and do the cleaner thing which is to eliminate anything with “highway=service”
@zelonewolf Any advice on how we could block plazas from import?
They are defined by highway = pedestrian and area = yes, which appears correct per the pedestrian tag notes.
Seems like blocking either of those tags independently might be an issue.
Hi!
Is this an ongoing discussion? I support the idea to add
[‘highway’ !~ ‘bus_stop’]
and
[‘public_transport’ !~ ‘platform’]
The waiting areas at bus stops are supposed to be tagged as highway=bus_stop. So there is nothing we can do to correct OSM data here.
As far as I can tell, there is no need to include highway=bus_stop. Even if there is a road, the bus stop is never part of the road. There are always extra nodes for the stop. So we wouldn’t exclude any actual streets.
Brilliant!
Yeah, I’m always looking to improve the OSM query.
Thanks! 
I had seen this locally, glad someone did the legwork to figure out why those were still coming in despite not being marked as streets
Well observed! At the bus station, a lot of “streets” will suddenly disappear from CS. 
To be completely honest, I still have an open question for @JamesChevalier though, because I had listed a few bus stops for myself that are no real streets, and when I now check in OSM, I notice they are not marked as bus stops but as platforms, and although [‘highway’ !~ ‘platform’] is already in the query, they exist in CS as streets. Why?
One example among others: Way: Dominicusstraße/Hauptstraße (392446581) | OpenStreetMap
At the bus station, a lot of “streets” will suddenly disappear from CS.
I don’t quite understand. Are there streets in OSM that aren’t present in CityStrides?
although [‘highway’ !~ ‘platform’] is already in the query, they exist in CS as streets. Why?
I continually improve that OSM query, so earlier cities brought in won’t be as “clean” as latter cities.
There are “streets” in CS that should not be streets and that will be eventually removed when you update CS.
At the bus station of my hometown, there are following streets in CS: Gate 07, Gate 08, Gate 15, Gate 16, Gate 17, Gate 18, Gate 19, Gate 20, Gate 25, Gate 26, Gate 27, Gate 28, Gate 29, Gate 30, Gate 31, Gate 32, Gate 33.
As you can guess, these are no real streets but platforms where the long distance buses stop.
Thanks for the quick answer.
@r.weide brought this up earlier. I am just supporting his idea, because I noticed a looot of bus stops in my city, that are “streets” or part of streets in CS. 
@dbafounta: [‘highway’ !~ ‘bus_stop’] seems more general than [‘highway’ !~ ‘platform’] , as it includes the bus stops, that do not have a platform. The platform then is added with public_transport=platform, if there is one.
Hi All, new to the message boards here, but I think this question is most relevant to the thread here. Is there a recommended way to handle a single “street” that is a mix of inaccessible expressway sections and other easy to run segments?
The Mystic Valley Parkway in Medford, MA has quite a few nodes along different stretches, but the segments in the screenshot are not accessible safely. Inaccessible sections all appear to be marked as highway-trunk OSM Link or highway-trunk_link OSM Link 2.
From my reading of the google doc, these should be ignored in CS. Is there a way to push an update and get them removed? Or best to just mark manually complete?
Any advice is much appreciated!
Those segments should be marked as foot=no in openstreetmap.
Thanks, Brian. That was quite an easy fix. Appreciate your help here.
Although we have [‘foot’ !~ ‘no’], please could you add [‘foot’ !~ ‘private’] and [‘foot’ !~ ‘customers’] to match the equivalents in the query for access=? Having people add an overly permissive access= tag to a highway=pedestrian way could be unhelpful for routing software.
There is Google doc of what parameters denote a street to import or to exclude? I have been looking for something to reference, is the link to that doc available to share?
I’m confused about the exact tagging…
When I view Key:foot - OpenStreetMap Wiki it points me to Key:access - OpenStreetMap Wiki as the list of all possible values.
So I don’t know if I’m filtering on the foot=customers tag or the access=customers tag. Similarly, the foot=private / access=private distinction is confusing.
I was informed by a an OSM moderator of Vermont that when I changed a bunch of “Private” roads to access=private, that removes the road from routing engines. I validated the claim as accurate. We should only label any road as access=private if there is a sign forbidding access/trespassing
I then changed all of these roads back to access=yes but changed from highway=residential to highway=service highway=driveway
