Artificial Wasteland artwaste.land

The measured dimensions of catalogued paintings

7 of 7 conditions 720,147 rows · 28.6 MB · 10 files

720,147 rows: every catalogued painting in four open sources whose picture-surface dimensions could be read out of the catalogue's own prose, from 736,698 candidate records across four sources.

The files

FileRowsBytessha256
paintings-aic.csv 3,526 256,973 34a3dba703fdd17377238d7f5470d6a9858ff6394f0bd3d6c80a5f0f7fad2600
paintings-met.csv 8,324 647,726 54a6ec0449ef744d5b15a150233279731842315be9f301e908130e9a50c3b95a
paintings-nga.csv 4,178 246,425 48207bf38f3929bb382b2e3c6e5f1829a8395ed5a617907dc275e51128abbb12
paintings-wd-1.csv 240,000 9,835,954 14e3e8c56a36ac6dc9031a2952349068265cba58ee726453530728fbe9db59d8
paintings-wd-2.csv 240,000 9,793,468 64bd8096575fa6c129a2fc14b171fd316a4e8b86f219da5729c1530b149d3eba
paintings-wd-3.csv 224,119 7,829,486 788160cbee6e35901242a0edfb21dae03fac39d5986cf6a47071eccceed2d4d6
README.md n/a 3,906 b7ab2d891a1e3fe33a6c01ff32fdffa9044d1a7ce9de41b22efda0f8e9ee8c27
schema.json n/a 4,543 35365cdb724e51660f42227e3f6a519cf2cafe9a88ac3b32bdc80d16ba2bf43e
sources.json n/a 4,815 9d8ee7280fa8c126d2fb82937d87d294a194930789cddabc7aad728a2c033c1e
validate.mjs n/a 15,198 67aceadfaea1f0b4611fa24f550c1b648437f25fbf20f19bf864292700aa69a2

Take the folder and check it yourself: node validate.mjs in the downloaded directory re-runs every row against the schema and every file against the digest above, with no dependencies and no network. The dataset's own full account is README.md.

A row, field by field

Origin is the column that decides whether we may give a field away at all. built is ours; redistributed is a source's own content, and it may only ship where that source's licence permits it.

FieldTypeOriginWhat it is
catalogue string built Which catalogue: aic Art Institute of Chicago, met The Metropolitan Museum of Art, nga National Gallery of Art Washington, wd Wikidata. THREE OF THESE ARE SINGLE MUSEUMS AND THE FOURTH IS A CROWD-SOURCED DATABASE that supplies 97.8 per cent of the rows. Never pool them.
height number redistributed Picture-surface height, in `unit`, as the catalogue printed it.
width number redistributed Picture-surface width, in `unit`.
unit string or null redistributed cm or in. Null on every Wikidata row: the query returns truthy values without the unit qualifier, so 704,119 rows have no recorded unit and their absolute sizes are not comparable across rows.
ratio number built Long side divided by short side, so a portrait and a landscape of the same proportions are one shape. Ours, and unit-free, which is why it is the only field comparable across all four catalogues.
in_tool boolean built Whether this row is inside the 1:1 to 3:1 window the /tools/canvas-ratio/ payload encodes. False on the 8,592 rows longer than 3:1, which the instrument never searches.
frame_height number or null redistributed Framed height where the catalogue printed one separately.
frame_width number or null redistributed Framed width where printed separately.
canvas_height number or null redistributed Canvas or support height where printed separately from the picture surface.
canvas_width number or null redistributed Canvas or support width where printed separately.
year_from number or null redistributed Earliest year of the catalogue's date range for the work.
year_to number or null redistributed Latest year of that range.
department string or null redistributed The museum's own department label, where it has one. Null throughout Wikidata.

Where it came from, and how we know we may pass it on

The dimension parse, the ratio, and the merged record set

Artificial Wasteland · https://artwaste.land/tools/canvas-ratio/

Licence
CC0 1.0 Universal
How we know
publisher statement
Accessed
2026-08-01
Covers
Every field. The numeric values within the redistributed fields originate in the four catalogues below.

The parse and the merged record set are ours and released CC0. The dimension NUMBERS inside it come from the four catalogues named below, every one of which is itself CC0, so no row in this file carries a more restrictive term than CC0. That is why one composite source covers the redistributed fields: the claim is true of every row, not of the file on average.

Art Institute of Chicago public API

Art Institute of Chicago · https://api.artic.edu/api/v1/artworks

Licence
CC0 1.0 Universal (all fields except `description`, which we do not use)
How we know
publisher statement
Accessed
2026-08-01
Covers
Rows where catalogue = aic.

The API states its own terms in every response. Re-read 2026-09-04 at https://api.artic.edu/api/v1/artworks?limit=1, which returns: "The `description` field in this response is licensed under a Creative Commons Attribution 4.0 Generic License (CC-By) and the Terms and Conditions of artic.edu. All other data in this response is licensed under a Creative Commons Zero (CC0) 1.0 designation and the Terms and Conditions of artic.edu." This dataset uses dimensions and dates, never `description`.

The Metropolitan Museum of Art Open Access (MetObjects.csv)

The Metropolitan Museum of Art · https://github.com/metmuseum/openaccess

Licence
CC0 1.0 Universal
How we know
licence file
Accessed
2026-08-01
Covers
Rows where catalogue = met.

The repository's LICENSE file was fetched 2026-09-04 from https://raw.githubusercontent.com/metmuseum/openaccess/master/LICENSE and is the CC0 1.0 Universal legal text. A copy is committed at research/data-room/licences/MET-OPENACCESS-LICENSE.txt.

National Gallery of Art Open Data Program (objects.csv)

National Gallery of Art, Washington · https://github.com/NationalGalleryOfArt/opendata

Licence
CC0 1.0 Universal
How we know
licence file
Accessed
2026-08-01
Covers
Rows where catalogue = nga.

The repository's LICENSE file was fetched 2026-09-04 from https://raw.githubusercontent.com/NationalGalleryOfArt/opendata/main/LICENSE and is the CC0 1.0 Universal legal text. A copy is committed at research/data-room/licences/NGA-OPENDATA-LICENSE.txt.

Wikidata Query Service (P31=Q3305213 with P2048 and P2049)

Wikimedia Foundation and the Wikidata contributors · https://query.wikidata.org/sparql

Licence
CC0 1.0 Universal
How we know
publisher statement
Accessed
2026-08-01
Covers
Rows where catalogue = wd, 97.8 per cent of the file. Crowd-sourced, not a museum.

Wikidata:Licensing, read 2026-09-04 at https://www.wikidata.org/wiki/Wikidata:Licensing: "All structured data in the main, property and lexeme namespaces is made available under the Creative Commons CC0 License (Public domain)." The dimensions used here are main-namespace structured data.

Required attributionDimensions from the open catalogues of the Art Institute of Chicago, The Metropolitan Museum of Art and the National Gallery of Art, Washington, and from Wikidata, all released CC0. The parse that decides which printed numbers are the picture surface is by the Artificial Wasteland.

Before you compute anythingThe four catalogues are counted separately and must never be pooled. Wikidata contributes 704,119 of 720,147 rows (97.8 per cent); the three museums together contribute 16,028. A statistic computed over the whole file is a statistic about Wikidata.

The seven conditions, and what was measured

  1. Bytes 6 data file(s), 28,610,032 bytes, largest 9.4 MiB
  2. Shape 720,147 rows checked, 0 violation(s)
  3. Count 720,147 rows across 6 file(s), all counts agree
  4. Fixity 10 file(s) verified by sha256
  5. Provenance 5 source(s), each with a licence from the table
  6. Origin 13 field(s): 3 built here, 10 redistributed under a licence that permits it
  7. Rebuild research/data-room/emit/canvas-proportions.mjs (byte-equality proved by: node research/data-room/publish.mjs --check)

← every dataset, and the three we could not publish