The dataset
The raw GPS archive and derived delay tables behind this site — every train we track, not passenger services alone — recorded continuously since 28 July 2026: available for research, journalism and product work.
What's inside
- Raw archive — every payload from the live feed, stored append-only exactly as received, before any parsing.
- Position pings — cleaned GPS observations per train: timestamp, latitude/longitude, speed, live flag.
- Station events — arrivals and departures per train, per service date, per station, with the scheduled time, the delay in minutes, and a provenance field (operator record or GPS).
- Daily train stats — per-train rollups: average delay, on-time percentage, day-of-week and per-station breakdowns.
- Reference tables — trains, stations and versioned timetables (a schedule change is a new version, not an overwrite).
Historical dataset
$149 one-time
A full export of the archive and derived tables to date, in documented CSV/JSON. Need a recurring feed, a specific slice, or a commercial licence? Ask for a custom quote.
Good to know
- Collection started on 28 July 2026 and coverage grows daily — email for current row counts before you buy.
- The data describes train movements only — there is no personal data of any kind in the dataset.
- How the archive is collected, and what it does and does not measure, is set out on how this site works — worth reading before you buy.
- An independent archive, not a Pakistan Railways product. Data is provided as observed, without warranty — the full licence, including what a purchase does and does not let you do with it, is in the dataset terms.