[Talk-transit] Uploading public transport data on OSM

Wiklund Johan johan.wiklund at entur.org
Fri Jan 19 10:40:36 UTC 2018

But then how would an outsider know what the operator code or ID is, and where this operator+id is unique? If GTFS data is the source of your OSM stops, I would say a dataset reference is far more relevant than operators or network since the ID is unique in a GTFS dataset, but only in that dataset.

Overlapping datasets always cause problems of course, but only if they are not properly referenced.

Johan Wiklund
Data manager 
johan.wiklund at entur.org


-----Original Message-----
From: Andrew Davidson [mailto:theswavu at gmail.com] 
Sent: fredag 19. januar 2018 04.12
To: talk-transit at openstreetmap.org
Subject: Re: [Talk-transit] Uploading public transport data on OSM

On 19/01/18 01:32, Stephen Sprunk wrote:
> Agreed; ref:gtfs just won't work, and ref:OPER probably would.

I had always thought that ref was the public facing reference that is used (ie: what's on the bus stop sign) and in the GTFS scheme this is stop_code. The stop_id is supposed to be unique and is not necessarily the same as stop_code. Looking at taginfo the most common way to tag stop_id seems to be gtfs_id.

So if there is more than one stop_id for a stop (ie: appears in more than one GTFS feed) I would have thought something like:


I have read elsewhere that some mappers prefer to use NETWORK, rather than OPERATOR, because their systems get put up for tender for operation on a regular basis so NETWORK is more persistent.

I was wondering if there are more than one stop_id is it worthwhile tagging something like this:


So that the data consumers will find multiple values under gtfs_id and know to look for an operator or network value. Or is it better to leave it blank? I guess it would depend on whether or not you have put qualified gtfs_id's on everything or just the shared stops.

Talk-transit mailing list
Talk-transit at openstreetmap.org

More information about the Talk-transit mailing list