Datasets:

idle-intelligence
/

metar-stations

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
text: list<item: string>
  child 0, item: string
n_networks: int64
n_stale_dropped_from_wider_window: int64
n_roster_total: int64
window_days: int64
window_start: timestamp[s]
n_stations_all: int64
n_active_total: int64
n_dropped: int64
build_date: timestamp[s]
to
{'build_date': Value('timestamp[s]'), 'window_start': Value('timestamp[s]'), 'window_days': Value('int64'), 'n_networks': Value('int64'), 'n_roster_total': Value('int64'), 'n_active_total': Value('int64'), 'n_dropped': Value('int64'), 'n_stale_dropped_from_wider_window': Value('int64'), 'n_stations_all': Value('int64')}
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 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, 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 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, 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/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, 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
              text: list<item: string>
                child 0, item: string
              n_networks: int64
              n_stale_dropped_from_wider_window: int64
              n_roster_total: int64
              window_days: int64
              window_start: timestamp[s]
              n_stations_all: int64
              n_active_total: int64
              n_dropped: int64
              build_date: timestamp[s]
              to
              {'build_date': Value('timestamp[s]'), 'window_start': Value('timestamp[s]'), 'window_days': Value('int64'), 'n_networks': Value('int64'), 'n_roster_total': Value('int64'), 'n_active_total': Value('int64'), 'n_dropped': Value('int64'), 'n_stale_dropped_from_wider_window': Value('int64'), 'n_stations_all': Value('int64')}
              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.

METAR Stations (Active, 7-Day Window)

A worldwide list of METAR/ASOS weather stations, sourced from the Iowa Environmental Mesonet (IEM) ASOS/AWOS networks. stations.json keeps only stations that actually reported at least once in the trailing 7-day window; stations_all.json keeps the full roster, including stations that were silent over that window (some other station lists still carry these as open), with a flag for whether each one reported.

Files and fields

stations.json, active stations only, compact rows for a browser client, one array per station:

field	type	unit
icao	string	station id as IEM uses it: the 4-letter ICAO code, or a 3-character FAA id for 2,444 stations, mostly in the US
lat	float	degrees, 4 decimals
lon	float	degrees, 4 decimals
elev_m	float	meters
name	string	station name
country	string	ISO-2 country code

stations_all.json, every roster station whether active or silent, same row shape as stations.json plus two trailing fields:

field	type	unit
icao	string	station id, same convention as `stations.json`
lat	float	degrees, 4 decimals
lon	float	degrees, 4 decimals
elev_m	float	meters
name	string	station name
country	string	ISO-2 country code
reported	bool	true if the station had >=1 tmpf reading in the 7-day window
last_report_utc	string or null	last confirmed report in the pulled window, or null if `reported` is false

stations.parquet, the same rows as stations.json plus two fields used to build the list, for the Hub's table viewer:

field	type	unit
icao	string	station id as IEM uses it: the 4-letter ICAO code, or a 3-character FAA id for 2,444 stations, mostly in the US
lat	float	degrees, 4 decimals
lon	float	degrees, 4 decimals
elev_m	float	meters
name	string	station name
country	string	ISO-2 country code
last_report_utc	string	ISO 8601 timestamp, last confirmed report in the pulled window
source_network	string	IEM network id, e.g. FR__ASOS

How it was built

build_active_stations.py walks every IEM network whose id ends in ASOS (one network per country or per US state/Canadian province), pulls each network's station roster from IEM's geojson endpoint, then pulls each network's temperature (tmpf) archive over a 7-day window from IEM's asos.py endpoint. A station goes into stations.json only if it has at least one non-empty tmpf reading anywhere in that window; a station's presence in IEM's roster feed (online=true) is not treated as evidence of activity, since some roster entries stay marked online for years after they stop reporting. stations_all.json is built from a second pass over the same per-network geojson roster endpoint (kept for every station this time, not just the active ones) with the reported flag and last_report_utc taken from the already-fetched activity data, so the archive endpoint is never queried twice for the same network.

IEM lists a depth or a wrong negative elevation for 17 coastal and offshore stations (oil platforms, islands, lighthouses); the script sets these to 0 m. Stations that really are below sea level, around the Caspian Sea and in the Imperial Valley, keep their elevation.

Elevations are IEM's; 17 wrong negative values on active coastal and offshore stations are set to 0 m by the build, while silent stations keep IEM's values, a few of them wrong (helidecks, ships), which does not matter for estimates since only reporting stations are used.

Roster seen: 7,534 stations across 266 networks. Kept in stations.json (reported at least once in the 7-day window): 5,614. Silent in that window but present in stations_all.json: 1,920. Build date: 2026-09-28, window 2026-09-21 to 2026-09-28.

Source and license

Data pulled from the Iowa Environmental Mesonet (IEM), Iowa State University: https://mesonet.agron.iastate.edu/. IEM's data is public domain; attribution is appreciated (see https://mesonet.agron.iastate.edu/disclaimer.php). This dataset is released under CC0-1.0. IEM's own terms place no restriction on redistribution beyond that appreciated attribution, so CC0-1.0 is used here rather than a more restrictive license.

How to regenerate

venv/bin/python build_active_stations.py

The window is computed from the current date each run, so re-running later produces a different 7-day window and station set. Progress checkpoints in checkpoint_rows.jsonl/checkpoint_done_networks.txt (activity) and checkpoint_roster_all.jsonl/checkpoint_roster_done_networks.txt (full roster), so an interrupted run resumes per-network instead of re-fetching from scratch.

Using your own station list

The weather-web library reads any list in the same row format; see its README (github.com/idle-intelligence/weather-web).

Downloads last month: 301

Total file size:

1.09 MB