Misplaced subway stations in Paris #46

tclementdev commented 2026年02月26日 20:10:02 +01:00

The Opéra and Saint-Lazare subway stations are not at the correct locations in Paris: they seem to have merged with the Havre-Caumartin station.

The Opéra and Saint-Lazare subway stations are not at the correct locations in Paris: they seem to have merged with the Havre-Caumartin station.

zyphlar commented 2026年02月26日 22:59:48 +01:00

It looks like someone recently changed them in OSM, maybe combining them? CoMaps will create approximate locations for the "station" probably at the center, and isn't detailed enough to handle multi-station complexes, so the first question is whether OSM is correct (are these stations related like this?) and the second question is whether comaps handles this situation properly (probably not the best, but what could be better given the OSM data?)

https://www.openstreetmap.org/relation/7731534

https://www.openstreetmap.org/changeset/177761173

It looks like someone recently changed them in OSM, maybe combining them? CoMaps will create approximate locations for the "station" probably at the center, and isn't detailed enough to handle multi-station complexes, so the first question is whether OSM is correct (are these stations related like this?) and the second question is whether comaps handles this situation properly (probably not the best, but what could be better given the OSM data?) https://www.openstreetmap.org/relation/7731534 https://www.openstreetmap.org/changeset/177761173

tclementdev commented 2026年02月26日 23:20:10 +01:00

It's hard for me to tell whether the OSM data is correct. I think it's trying to indicate there are connections between the stations but that seems like information that could be used for routing, not for visually merging the stations on a map.

It's hard for me to tell whether the OSM data is correct. I think it's trying to indicate there are connections between the stations but that seems like information that could be used for routing, not for visually merging the stations on a map.

zyphlar commented 2026年02月27日 00:46:13 +01:00

I agree. It could be a difficult judgment call about the right way to represent things, both as data and visually in the app.

Does the routing work properly, is the main issue that the station location is inaccurate?

I agree. It could be a difficult judgment call about the right way to represent things, both as data and visually in the app. Does the routing work properly, is the main issue that the station location is inaccurate?

tclementdev commented 2026年02月27日 09:09:18 +01:00

I'm not sure how routing works. I would assume the connections between lines is a separate information from the GPS locations of the stations. If we can just use the proper locations without changing anything else, I guess we should be good? As of now, people looking to go to the stations using Comaps would find themselves at the wrong station.

I'm not sure how routing works. I would assume the connections between lines is a separate information from the GPS locations of the stations. If we can just use the proper locations without changing anything else, I guess we should be good? As of now, people looking to go to the stations using Comaps would find themselves at the wrong station.

fnetX commented 2026年03月13日 01:34:15 +01:00

Hi! I am currently visiting Paris, and I appreciate your work on this. (I also enjoy that RER lines just appeared during my visit).

Actually, I got lost in the maze of these subway stations just a few days ago. I don't exactly remember where I came from, but I left the metro at Opéra and was looking for metro line 12. According to CoMaps, it goes through the same station, but you actually walk about 1km underground, and I found no signs (because they probably didn't expect someone to leave there for this line). Actually, it would have been better to go to Madeleine (or Haussmann Saint-Lazare?) and take the 12 from there.

Looking at the available data, I think there is currently no clear intended way to differentiate "same-station" multi-modal access (e.g. same or similar name and close relation) compared to the Paris case of having stations somehow connected, but not really being the same station? I guess it would require adding additional tagging to OSM data to allow for that?

I like the visual representation in the official subway maps (such as https://assets-web.bonjour-ratp.fr/PLAN_REGION_2025_12_06_12_2025_adee59ba82.pdf). It is clear from the map which stations are "the same" (thick white lines) and which are distinct stations but somehow reachable (smaller white lines).

Hi! I am currently visiting Paris, and I appreciate your work on this. (I also enjoy that RER lines just appeared during my visit). Actually, I got lost in the maze of these subway stations just a few days ago. I don't exactly remember where I came from, but I left the metro at Opéra and was looking for metro line 12. According to CoMaps, it goes through the same station, but you actually walk about 1km underground, and I found no signs (because they probably didn't expect someone to leave there for this line). Actually, it would have been better to go to Madeleine (or Haussmann Saint-Lazare?) and take the 12 from there. Looking at the available data, I think there is currently no clear intended way to differentiate "same-station" multi-modal access (e.g. same or similar name and close relation) compared to the Paris case of having stations somehow connected, but not really being the same station? I guess it would require adding additional tagging to OSM data to allow for that? I like the visual representation in the official subway maps (such as https://assets-web.bonjour-ratp.fr/PLAN_REGION_2025_12_06_12_2025_adee59ba82.pdf). It is clear from the map which stations are "the same" (thick white lines) and which are distinct stations but somehow reachable (smaller white lines).