Hugging Face's logo

Datasets:
alreadyopen
/
sa-transit

Modalities:
Text
Geospatial
Formats:
parquet
Size:
1M - 10M
Tags:
geospatial
transit
gtfs
gtfs-realtime
adelaide
south-australia
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
Dataset Viewer
Auto-converted to Parquet Duplicate
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
booking_rule_id: string
booking_type: string
prior_notice_last_day: string
prior_notice_last_time: string
message: string
to
{'agency_id': Value('string'), 'agency_name': Value('string'), 'agency_url': Value('string'), 'agency_timezone': Value('string'), 'agency_lang': Value('string'), 'agency_phone': Value('string'), 'agency_fare_url': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2951, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 547, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/parquet/parquet.py", line 220, in _generate_tables
                  yield Key(file_idx, batch_idx), self._cast_table(pa_table)
                                                  ~~~~~~~~~~~~~~~~^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/parquet/parquet.py", line 156, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              booking_rule_id: string
              booking_type: string
              prior_notice_last_day: string
              prior_notice_last_time: string
              message: string
              to
              {'agency_id': Value('string'), 'agency_name': Value('string'), 'agency_url': Value('string'), 'agency_timezone': Value('string'), 'agency_lang': Value('string'), 'agency_phone': Value('string'), 'agency_fare_url': Value('string')}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

South Australia public transport

The Adelaide Metro General Transit Feed (buses, trains and trams in metropolitan Adelaide, plus the regional coach operators the same feed carries, such as Stateliner) as GeoParquet, JSON and vector tiles, rebuilt weekly by AlreadyOpen/sa-transit, with a snapshot of the GTFS-realtime feeds (vehicle positions, trip updates and service alerts) every 15 minutes. Every table, row and field of the source is kept.

File Holds
gtfs/<table>.parquet Every table of the feed (agency, routes, trips, stop_times, stops, shapes, calendar, calendar_dates, transfers, booking_rules, feed_info, and any table added later), every column as text, exactly as published
release-notes.txt The feed's own release notes
stops.parquet GeoParquet: every stop, every column, with a point geometry
route-shapes.parquet GeoParquet: one row per route and shape a trip uses: every route column, shape_id, trips, points, shape_dist_traveled, and the shape as a line
stops.json GeoJSON of every stop, every field
routes.json Every route, every field, with its shape_ids and trips
transit.pmtiles Vector tiles, layers routes and stops, zooms 4 to 14; every feature is in every zoom
index.json Row counts, the tables, and the feed's feed_info (version and validity dates)

Live feeds (bucket only)

On the bucket, refreshed every 15 minutes (GitHub may start a run late or skip one, so the history has gaps). Three GTFS-realtime feeds, each as <feed>:

<feed> Feed Daily Parquet: one row per
vehicles vehicle positions vehicle per snapshot (GeoParquet, a point per vehicle)
trip_updates trip updates (predicted arrivals and departures) trip per snapshot; stop_time_updates counts its stop predictions, which are in entity
alerts service alerts (disruptions, stop changes) alert per snapshot
File Holds
live/<feed>.json The latest feed decoded to JSON, every field, including the feed's own vehicle extension (air_conditioned, wheelchair_accessible as tfnsw_vehicle_descriptor on every vehicle descriptor)
live/<feed>.pb The same feed as received (protobuf)
live/<feed>-YYYY-MM-DD.parquet Every entity in every snapshot of that UTC day, with the main fields as columns and entity, the full entity as JSON

snapshot is the feed's own header timestamp, and the day is taken from it. On 2026-10-11 the service alerts feed's header timestamp was Adelaide local time written as if it were UTC (10.5 hours ahead during daylight saving), while the other two feeds were in true UTC; it is kept as published.

The bucket keeps the current month (plus the previous one until it is archived). The full history is on Kaggle, one dataset per feed and month, never changed once created: helenkwok/sa-transit-vehicles-YYYY-MM, helenkwok/sa-transit-trip-updates-YYYY-MM and helenkwok/sa-transit-alerts-YYYY-MM (e.g. https://www.kaggle.com/datasets/helenkwok/sa-transit-vehicles-2026-10).

manifest.json lists every file's sha256, the counts and the sources of the latest bake. Coordinates are WGS 84 as published in the feed.

Sources and terms

Data from the Department for Infrastructure and Transport (Adelaide Metro), Government of South Australia: Adelaide Metro General Transit Feed and its GTFS-realtime API, licensed under Creative Commons Attribution 4.0. Changes: converted to Parquet, GeoParquet, JSON and vector tiles, with shapes joined to routes; nothing removed. Provided as is, with no warranty, not endorsed by Adelaide Metro or the Government of South Australia. Do not rely on it for travel; use Adelaide Metro's own services.
Downloads last month
-
Number of rows:
2,119,981
Total file size:
39.1 MB