Community Council Areas missing from OSM (17 of 46):
Colinton
Corstorphine
Craigleith/Blackhall
Craiglockhart
Cramond and Barnton
Drum Brae
Fairmilehead
Firrhill
Grange/Prestonfield
Kirkliston
Marchmont & Sciennes
Merchiston
Morningside
Murrayfield
Ratho and District
Silverknowes
Queensferry and District
This is awesome! Thanks for doing this. I’m curious if there’s a way now to create new nested cities so that gigantic areas (like Queens or Brooklyn) can be broken up into more bite sized chunks.
So what do you see this as the scope of this? Limited to fixing double street counts, or can it be expanded to split big cities into smaller chunks or combine small villages into bigger targets?
All of the city border data comes from OpenStreetMap. If sub-sections of a city exist in there, I can import those into CityStrides as nested cities.
I’ve done that for a number of larger cities & I’m happy to add more.
One thing I’ve noticed is some roads in “nested cities” that are in the nested city but not the larger city don’t count at all. I don’t know why, I haven’t investigated in OSM, but for instance in my recent activity (CityStrides) there are 15 unique street names in my completed list, but it only says 9 completed (the 9 that would previously have double-counted). Before Benfica was nested in Lisboa, this run would have shown 24 completed streets (9 x 2 + 6) but the solution to that problem now under-counts the streets I did complete
I’ve been trying to figure out Aberdeen, actually. The smaller nested cities has been helpful to me since I live in the midst of all those, but it’s clearly a broken system. I covered a road last week (Prospect Terrace) which has the same name as a road in Dyce. OSM/CS seems to treat them as the same road, given that Dyce is part of Aberdeen (despite being significantly far apart). I wonder if mapping Dyce as a separate, nested boundary would help sort problems like that out?
I also couldn’t figure out why “Castlehill and Pittodrie” or “Froghall, Powis and Sunnybank” are a separate area of the city while Cults, Dyce, Bieldside, Kincorth, etc. aren’t. To my knowledge, these aren’t historically separate towns which merged (unlike Torry, Old Aberdeen, and Bridge of Don, but are somewhat random groupings. They don’t seem to follow any specific metrics for being bound - any insights as to what sets them apart?
I think the boundaries for the nested ‘cities’ in OSM for Aberdeen are, like Edinburgh, for community councils, but I think it may be a work in progress and not all of the Aberdeen community council areas have been added to OSM yet. Here’s the full list: Community Councils | Aberdeen City Council
Yeah. From the perspective of Aberdeen, Prospect Terrace exists twice: Prospect Terrace - CityStrides
OSM does not have knowledge of “streets”, it only knows about ways. A Way is a collection of Nodes, and a single Way may be the entirety of a Street or it may be a section of it. Each Way doesn’t know about another Way, so within OSM there’s no possibility (that I’m aware of yet) to join up multiple Way records to determine its full street.
So, when I’m importing streets from OSM I have to figure out how to join up all these individual way records into what we all collectively know as the Street. I decided to do that based on the street name.
I expect that you completed the street within the nested city (which now doesn’t count towards your total count), but that the street in the parent city is either much larger (new city imports now cut off nodes at that city’s border) and/or the same issue and it’s actually multiple streets.
Take Rua E for example.
You completed it within Benfica: CityStrides
But in Lisboa, there a four Rua E to complete: Rua E - CityStrides
I’m going to think over the idea of switching my calculations in reverse.
Instead of not counting streets in the nested city … maybe I should not count the streets in the parent city.
My immediate concern there, however, is that there are many cases of the nested city not existing (in OSM or in CityStrides).
Even if a city is “fully nested”, counting streets in the nested cities would lead to double counting since it’s common for streets to mark the border between districts/suburbs. In my view having the occasional undercount because of repeated street names (possibly solvable) is the lesser evil compared to guaranteed overcounting due to summing-up at nested level.
Hi and thanks for CityStrides! I’m new and was recently thinking wouldn’t it be cool to see all the streets I have been running and I found this. Fantastic!
My home city Stockholm (Stockholms kommun, Stockholms län - CityStrides) has 4185 streets and it would be really nice if nested cities could be added for the 13 or 14* city districts Stockholm is divided into.
*) To complicate things a bit just recently on 1 July “Hägersten-Liljeholmen” and “Älvsjö” were merged into “Hägersten-Älvsjö”, but that has not been updated on OSM yet. I don’t know if any regards to that should be taken.
I’m not sure if mine are being counted twice or not. They’re certainly all listed twice. For example, the walk I did yesterday says 39 streets completed. Counting the streets listed under this tab myself, there are 94. 55 of these are unique. So I’m not sure where the 39 comes from?
Also most the duplicates appear under completed once I’ve got all the nodes, but sometimes it will appear once under completed and once under progressed, though all the nodes are green. Does this mean I’m only getting credit in either Islington OR Greater London, but not both?