Hugging Face's logo

Datasets:
Bekzod25
/
terra-atlas

Modalities:
Tabular
Text
Formats:
parquet
Languages:
English
Russian
Size:
100M - 1B
Tags:
history
archaeology
geography
gazetteer
radiocarbon
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
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
source: string
license: string
tier: string
retrieved: string
id: string
kind_of: string
name: string
name_ru: null
constellation: string
path_ra_dec: string
ra_deg: null
dec_deg: null
mag: null
dist_pc: null
spect: null
-- schema metadata --
terra_atlas_kind: 'sky_object'
to
{'source': Value('string'), 'license': Value('string'), 'tier': Value('string'), 'retrieved': Value('string'), 'id': Value('string'), 'name': Value('string'), 'name_ru': Value('string'), 'value': Value('float64'), 'unit': Value('string'), 'uncertainty': Value('float64'), 'domain': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 149, 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 129, 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 489, 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 2818, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2355, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2380, 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/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 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              source: string
              license: string
              tier: string
              retrieved: string
              id: string
              kind_of: string
              name: string
              name_ru: null
              constellation: string
              path_ra_dec: string
              ra_deg: null
              dec_deg: null
              mag: null
              dist_pc: null
              spect: null
              -- schema metadata --
              terra_atlas_kind: 'sky_object'
              to
              {'source': Value('string'), 'license': Value('string'), 'tier': Value('string'), 'retrieved': Value('string'), 'id': Value('string'), 'name': Value('string'), 'name_ru': Value('string'), 'value': Value('float64'), 'unit': Value('string'), 'uncertainty': Value('float64'), 'domain': 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.

TERRA Atlas of Historical Data

A warehouse of normalized extracts from open historical datasets for the TERRA civilization-simulation project (an epoch-travel game aiming at historically honest scenes). This is not a scrape of the web: every record comes from a published open dataset and carries provenance.

Record contract

Every row in every parquet file has four mandatory provenance fields:

field meaning
source URL or stable identifier of the origin record/dataset
license license of the origin dataset (see per-source table below)
tier confidence tier: K = known from a source (document, find, database record); T = typical for the epoch/place (norms, no direct record); R = reconstruction / model output, honestly flagged
retrieved ISO date the record was fetched

Record kinds (stored in parquet key-value metadata terra_atlas_kind): person, place, polity_snapshot, event, artifact, grid_layer, source (dataset-level provenance row).

Layout: data/<source>/<kind>_*.parquet (snappy, shards ≤ 256 MB). manifest.json documents per-source status, chunk checksums and sizes.

Sources

source what kinds license tier
pleiades ~42k places of the ancient world: coordinates, feature types, time periods, location precision place CC-BY-3.0 K
p3k14c ~180k archaeological radiocarbon dates (lab id, age BP, sd, coordinates, site, country) — settlement-density proxy. Year field is 1950 − age_bp, uncalibrated event CC-BY-4.0 K
dplace D-PLACE Ethnographic Atlas: ~1.3k societies (coordinates, focal year) + long-format variable values with code labels polity_snapshot CC-BY-4.0 K
seshat Seshat Databank, official Equinox packaged data (Zenodo mirror 10.5281/zenodo.6629022) polity_snapshot AGPL-3.0 per Zenodo record; Seshat site states CC-BY-NC-SA-4.0 — do not ship without checking K
met The Met Open Access: ~490k object metadata rows (title, begin/end year, culture, medium, public-domain flag). No images; api_url returns primaryImage artifact CC0-1.0 K
naturalearth Natural Earth 10m vectors: countries, populated places, rivers, lakes; geometry as GeoJSON strings + bbox + representative point place Public Domain K
wikidata_rulers First vertical slice “Mesopotamia −2500…−500”: rulers (reign dates, positions) and kingdom capitals with coordinates. Full pre-1500 persons layer is registered as a dump-scale plan, not fetched person, place CC0-1.0 K
hyde HYDE 3.3 (Utrecht University mirror): test slice — population-count grid (popc) for 2000 BC, 5 arcmin, zero cells omitted; remaining slices listed as pending URLs in the manifest grid_layer CC-BY-4.0 R (model reconstruction)

Honest-uncertainty policy

The atlas marks, rather than hides, how much is actually known: prehistory has no names or conversations — its ceiling is archaeology (tier K for material finds), everything else is T/R by definition. Model grids (HYDE) are tier R even though they are the best available estimates. Downstream consumers (scene compilers, NPC dossiers) must show provenance, e.g. “source: Pleiades 912985” vs “reconstruction by analogy”.

Reproduction

Collection pipeline (resumable, disk- and time-budgeted) lives in the TERRA repo: atlas/ package — python3 -m atlas.ingest --budget 1500, then python3 -m atlas.doctor, then HF_TOKEN=... python3 -m atlas.upload_hf.

Licenses

Each record carries its origin license; the compilation adds no restrictions. For shipping inside a game use only CC0 / Public Domain / CC-BY rows (mind CC-BY attribution) and exclude seshat until its terms are clarified.
Downloads last month
267
Number of rows:
181,311,114
Total file size:
1.18 GB