Just coming back to this again Brian would you be able to give an idea of what the algorithm is. ?
The following two stops are an inbound/outbound pair on a light rail line in Dublin:
stop_id,stop_name,stop_lat,stop_lon
"822GA00356","Kylemore","53.3265889562716","-6.34376337103717"
"822GA00357","Kylemore","53.3266344972522","-6.34380664554678"
Google maps represents these as one, the departure list in the attached image has both inbound and outbound trams: The 2 stops are located about 10m apart. I'm presuming google maps always uses GTFS and is using the stop file with the above extract. I'm wondering for the matching algorithm to work do stop names have to be exactly the same and be within a certain distance of each other ? Or is distance and name on its own enough ? Or is there an application where GTFS stops can be manually paired. Any light on this would be great ? Brian ?
If the pairing requires the name to be the same, then it is not ideal for other GTFS consuming applications that don't show paired stops in the same way. For these the stop names would be better represented as something like "Kylemore Inbound" and "Kylemore Outbound".
Combining stops like this also requires the combining of trips. It is as if a parent station stop is created on the fly with the 2 stops as children... could this be what the google algorithm does in a pre-processing step once it decides 2 stops are a pair ?
A