Rail data in GTFS format

[Hussein-Mahfouz]

Sources of rail data

Light Rail

Traveline (light rail) - link

• Dataset of scheduled bus, tram, and light rail services
• format: TransXChange
• updated: routinely updated every Monday – Thursday overnight
• UK2GTFS conversion: https://itsleeds.github.io/UK2GTFS/reference/transxchange2gtfs.html

Heavy Rail

ATOC (heavy rail) - link

• Full Timetable File - This contains details of all national rail passenger train services, associated shipping and bus links in a CIF format and is supplemented with passenger train service identifiers.
• format: CIF
• updated: it is only refreshed on an infrequent basis. It looks like it was last updated in 2011 - link
• UK2GTFS conversion: unclear
  • possibly the nptdr2gtfs() function. The nptdr.R script has a couple of function that are not documented (are they internal functions?)

Network Rail - link

• timetables and live information
• format: CIF and JSON
• updated: daily
• UK2GTFS conversion: https://itsleeds.github.io/UK2GTFS/reference/nr2gtfs.html

[Hussein-Mahfouz]

Data shared by Malcolm

1. ATOC: UK2GTFS/ATOC/GTFS/2023-05-03.zip (I thought ATOC was last updated in 2011?)
2. Network Rail: UK2GTFS/Network Rail Schedule/toc-full.CIF
3. TransXchange: UK2GTFS/TransXchange/GTFS/GTFS_2023_05_03

Questions for Malcolm

• Are all of these for rail only?
• Is (3) different to (1) or are they the same dataset?
• What would you recommend using to get the most complete dataset for (heavy + light) rail in England? I would like to create a GTFS feed and add it to the bus gtfs feed downloaded from BODS

Hi @mem48, could you help me with this issue please?

[mem48]

Transxchange is for bus and light rail, the ATOC data is still updated regularly

[mem48]

1 and 3 are different, but I periodically grab all data sources so tend to be on same/similar dates

[mem48]

I believe BODS now has light rail, but you would have to check. For heavy rail ATOC is the best source

[mem48]

See https://itsleeds.github.io/UK2GTFS/articles/ATOC.html

[Hussein-Mahfouz]

Thanks @mem48!

TODO:

• [x] download ATOC data
• [x] convert to gtfs
• [x] compare service patterns with bus gtfs data (and edit if necessary)
• [x] upload to inputs release
• [x] clip to study area boundary

[Hussein-Mahfouz]

Are some agencies missing from ATOC data?

I managed to download the data from ATOC and convert it using UK2GTFS::atoc2gtfs. I then clipped it to the study area (west yorkshire) using gtfstools::filter_by_sf(). The function has a keep argument which keeps any route as long as one of its stops is in the geometry you are clipping by.

After checking the agencies in the filtered gtfs, they are only 5 (down from 71). This makes sense except LNER is no longer available (LNER is in the agency file of the unfiltered feed: agency_name = "London North Eastern Railway", agency_id = "HB". These are the agencies remaining in the filtered feed:

I thought it was a bug in filter_by_sf(), but when I checked the routes table in the unfiltered feed, there are no LNER routes (routes with agency_id == "HB")

I tried another GTFS feed from Malcolm's onedrive UK2GTFS/ATOC/GTFS/2023-05-03.zip, but the same problem exists: LNER is in the agency table, but not in the routes table

@mem48 have you come across this issue before with atoc data?

[mem48]

Could they be in some other agency ID?

[mem48]

The raw data does not have the agency ID values they are added by UK2GTFS

[mem48]

Perhaps GR, East Coast? This may be a bug in UK2GTFS as East Coast was the old operator before LNER

[Hussein-Mahfouz]

@mem48 yes they are all under GR, East Coast (131 different entries in the routes table of the unfiltered feed, 18 entries in the filtered feed - see screenshot). Thanks! Is this worth opening an issue in UK2GTFS for?