Ottawa double street count

Just looked at " Most Completed Streets In The Last Week" and noticed that Ottawa is best place to run, cause streets are counted twice. An unfair advantage :grinning:

This city should probably be deleted.

No, it’s that there are two Ottawas for some reason.

Real city Ottawa, Ontario - CityStrides from Nominatim Demo

“Old” city from Nominatim Demo

I probably have to delete the “old” city. :man_shrugging:

Not sure if it’s just a question of deleting old city

is all part of

and there might be more

Yeah, there are some places like Ottawa where the Big City is admin level 6 - Nominatim Demo - and there are other Little Cities within it at admin level 8 e.g. Goulbourn Nominatim Demo

:man_shrugging: I’m more inclined to keep the admin level 8 places alongside the admin level 6 places, allowing “duplicate” streets.
I get requests here and there “hey why isn’t [some massive city] in CityStrides?” and “hey why isn’t [some little neighborhood] in CityStrides?” … it’s impossible for me to please everyone, so I try to refer to some external source e.g. wikipedia which states “Ottawa is the capital city of Canada.”

I’m also very uninterested in the conversation behind “what is a city?” :laughing:


Greetings everyone! A few of us here in Ottawa only recently found out about CityStrides so we’ve jumped on the wagon. Ottawa went through a significant change in 2001 when 11 municipalities combined to become the City of Ottawa. This would explain the “Old” Ottawa that appears in OSM.

From your perspective, should the “Old” Ottawa be deleted from CityStrides?

Deleting (Old) Ottawa from CityStrides and keeping only the amalgamated Ottawa would likely also mean deleting the other 11 municipalities that form the greater Ottawa area (Vanier, Rockcliffe Park, Gloucester, Nepean, Kanata etc…).

I personally like the idea of knocking off each of the former municipalities one at a time and find that contributes to the fun and challenge of CityStrides. I do agree with the original poster though that double counting the streets in the street count isn’t fair.

Ottawa is by no means unique with the “Old” counterparts. Even Canada’s largest city (Toronto) has the same setup… Not sure if streets are being double counted there though…

“Old” Toronto

Amalgamated Toronto Toronto, Ontario - CityStrides

There are two separate topics here:

  1. The ‘old vs new’ decision…
    Should we delete in favor of Toronto, Ontario - CityStrides
    Should we delete in favor of Ottawa, Ontario - CityStrides

  2. The ‘parent vs child’ decision…
    (I’m purposefully throwing a lot of terminology out the window for this one. I don’t care what people define a “city” as, or “neighborhood”, or “suburb”, etc - there are too many varying terms around the world to keep up for the purpose of this particular conversation. For this conversation it’s just a city and its components - a simple parent/child structure.)
    For most cities, it’s not worth bringing their components into CityStrides. This is because most cities are on the smaller side, and the next “layer down” is something along the lines of a neighborhood.
    For some cities, Los Angeles is an extreme example, the strictly-defined city is massive & for the purposes here in CityStrides it can make sense to import its components.

This ‘parent/child’ topic was also brought up in The municipality problem and Manhattan is weird now

Anyway - I don’t think a decision on the first topic affects a decision on the second. We can opt to delete the ‘old’ versions of a city, and also opt to keep the ‘child’ data.

My take is that - if the city is big enough - it’s ok to have a single huge ‘parent’ city (Ottawa) and all of its next-level-down ‘child’ “cities” (e.g. Goulbourn) in CityStrides.

1 Like

I like the parent/child idea a lot, especially if they are visually grouped in the profile/city view so that it’s easy to see how progress in the child entities works through to the parent. What number of streets would you have as the cut-off?

Hmmm. I’d have to do a fair amount of work to be able to include inter-city categorization (country → region → city → street → node is all easy enough and fully built out, but city → city would be massive). So I don’t expect any sort of visual grouping any time soon. :smile:

There’s enough work for me that I’d wait for someone to complain about a city being too big, then look into whether or not there’s another layer down that can be imported.

As for a cut-off on when to do this - there wouldn’t really be one because determining what gets imported into CityStrides is a manual process (How to help with missing/broken cities/regions).

My basic workflow is to take a country (which I’ve seen always exist at admin level 2 in OSM) and determine what admin level its regions are at. A lot of the time, it’s 4. It can vary, though - even within the country itself. Then I’m determining what the general/typical admin level for cities are for those regions. I usually spot-check a few regions for e.g. admin level 8 (pretty typical) or 6. I’ll do some searches for the names that come up to see if places like wikipedia mention them as cities. If that looks “good enough” then I’ll import all those cities & run some code that figures out local people’s progress.

Then I just sit back and wait for complaints. :laughing:

1 Like

I can understand your reluctance to open this particular can of worms since the mapping and classification of ‘child’ areas is likely even more idiosyncratic in OSM than country/region/city :laughing:. The main practical downside of just adding children on top of parents would be Ottawa-style double counting for any city where it is implemented, no?

1 Like

Yep. This thread is the first time I can recall receiving a complaint about this double counting, though. So. :man_shrugging: I’m not sure I care about that, either, if it took 7 years for anyone to notice/say anything.
If/as more people bring it up, I’ll move it higher up my list. :+1:

Interesting question here, and one I’ve thought about a lot as all my runs in Calgary, Alberta seem to count twice. Only just realized that this parent-child issue exists here too; with the “Census Division No. 6, Alberta”, which includes a big chunk of the province and is considered a ‘city’.

Any solution I can consider (exhaustively checking parent-child issues, adding the city name each completed street is associated with, cross-referencing completed streets to a users total count to eliminate duplicates) seems super computationally challenging, with the only impact being on the “Completed Streets This Week” leaderboard.

I’d vote to leave things as they are!

1 Like

The most pragmatic way to get parity for the leaderboard is probably for anyone who thinks they have a chance of making it to have their project “doubled-up” as well. Even that might not work as some areas already get triple-counted, where will it end? :laughing:

Census Division No.6, Alberta should probably go on the “Missing cities” sheet to be removed since it was downgraded to boundary:census.

1 Like

Thanks @8f7162110d9eeaf907ab good suggestion. I’ve added it to the list.
Will actually be nice to get a clean list of streets completed and representative total instead of this doubling up.

No, please keep it!

I am running an every street project of the pre amalgamated Ottawa. It’s nice to know my percentage completion and I am unable to find any “old” maps of Ottawa, Ontario, Canada.

Just so I understand this topic, the double streets in Ottawa is because there are two parents or “cities” overlapping each other?

I second this. I hadn’t ever really paid much attention to the ‘completed streets this week’ leaderboard, and I hadn’t realised that was something that people were especially interested in (or competing on!).

To my mind, it is much more interesting to follow the the number of street/percentage streets completed in a given city. Also, the differences between streets in different cities seems to make the ‘streets this week’ leaderboard a bit of an uneven comparison. For example some streets are hundreds of metres (or even kilometres) long. Others are short cul-de-sacs, barely a dozen metres in length. At least if you’re comparing the progress in a given city, you know that the same set of streets is available to everyone.

A final comment for me on the ‘overlapping’ nature of some cities. One of my target cities is Greater London. Having approx. 40k streets means that it will be a real marathon effort (and I could see myself getting disillusioned and disheartened!), but having the individual London Boroughs as short/medium term targets is a real benefit,

Therefore, I’m in favour of leaving things as they are!

Another vote for keeping it as is! I know a number of us are working on “Old Ottawa” vs other suburbs/amalgamated towns that happen to be part of the overarching ‘parent’.

My streets have been double counting for the last few days also. Based on the thread, I think it is because there is a Woodstock, CT and a South Woodstock, CT? The South Woodstock could be deleted if needed as everything is included in Woodstock.

Also, on a separate note - there is a street that I ran and on the map it looks like I ran a few feet down the road into a field and leaves the road without the purple lines on it or the road as completed. I manually completed because I wasn’t sure what else to do…guessing it was just a GPS issue?

Thank you for this great tool!!