SBB PT skims issues

[panostsolerid]

Hi,

This is an issue somewhat related to a previous open one (here). I am trying to make sense of SBB skims for PT specifically (no issues with car skims). I have summarised my issues/findings below.

1. Do the travel times in pt_travel_times.csv.gz refer to the total travel time for an OD pair or just the in-vehicle time? The previous issue I have attached above seems to imply that it is the former. But at the same time the documentation in here says that it is the in-vehicle time from the first to the last stop.

2. Are the travel times in pt_travel_times.csv.gz (and also in the rest e.g. pt_egress_times.csv.gz and pt_access_times.csv.gz) normalised in any way e.g. by hour or by frequency of services for a specific time period? I have tried several different time periods, namely i) 07:00-08:00, ii) 08:00-09:00, iii) 07:00-09:00 and iv) 07:00-10:00.
   For i) and ii) I am getting x mins for an OD, for iii) i am getting almost 2x mins and for iv) i am getting almost 3x mins.

Initially, I thought that i would have to normalise by the number of hours for each specified period, but then we ran some comparisons with google maps with a colleague.

From those experiments, it appears that when we divide the travel time for an OD in pt_travel_times.csv.gz with the frequency of services in pt_frequencies.csv.gz, we are getting the same travel time that Google suggests for a station to station trip.

So does this mean that the travel times in pt_travel_times.csv.gz are indeed in-vehicle times and that we need to normalise them by the frequency (number of services) for that time period? Are the travel times in pt_travel_times.csv.gz the cumulative time from all services running in that OD during that period?

Any help would be greatly appreciated !

Thanks and best regards!
Panos

[steffenaxer]

@mrieser or @jfbischoff you both might be the best to answer this issue.