Source

Fitzwilliam Museum — Department of Antiquities (Egyptian)

Egypt, Sudan, Nubia · Predynastic through Coptic

729

Objects

Inscriptions

1,934

Images

Browse all 729 objects → Visit source →

Sample objects

Source details

Type: museum_collection
Access: rest_api_with_key
Stack: axiell-ciim
License: CC-BY-NC-ND 4.0 (metadata); per-image rights vary
Modality: mdc
ID system: Fitzwilliam accession (E.LOT.YEAR format for Egyptian)
ID tier: tier-2
Inscription richness: 8/10
Cleanliness: 7/10

Form breakdown

jewelry: 502
vessel: 65
papyrus: 56
other: 37
stela: 35
tool: 13
statue: 11
architecture: 9
seal: 1

Notes

## Discovered schema (Fitzwilliam Museum Collection API, surveyed 2026-05-23) ### Endpoint + auth - API base: https://data.fitzmuseum.cam.ac.uk/api/v1 - OpenAPI: https://data.fitzmuseum.cam.ac.uk/api/v1/docs (Swagger UI) - Auth: Bearer JWT via POST /api/auth/signup (free) then /api/auth/login - Stack: Axiell Collections + CIIM (Knowledge Integration). v1.0.0, July 2022. ### HEADLINE: Fitzwilliam breaks the museum pattern — Flag 2 is FALSE here MET, BM, BKM all modelled inscriptions as not-modeled. **Fitzwilliam models inscriptions explicitly with full hieroglyphic transcriptions.** Sample of 109 objects (80 confirmed Egyptian via accession prefix `E.*` + find spot + period) yielded: - **6 objects (5%) carry at least one inscription** - **25 total inscription entries** across those 6 objects - **24 entries with non-empty transliterated text** - Average 4 inscriptions per inscribed object Example — accession E.207.1900, stelae: **9 inscriptions on one object**. First entry: location=None, method=None, type='inscription', description='', transcription='' ### Sample (n=109) by form category - jewelry n= 33 confirmed-Egyptian= 32 with-inscription=1 total-inscr=1 - stela n= 21 confirmed-Egyptian= 20 with-inscription=4 total-inscr=22 - vessel n= 16 confirmed-Egyptian= 7 with-inscription=0 total-inscr=0 - architecture n= 16 confirmed-Egyptian= 1 with-inscription=0 total-inscr=0 - statue n= 12 confirmed-Egyptian= 9 with-inscription=0 total-inscr=0 - other n= 6 confirmed-Egyptian= 6 with-inscription=1 total-inscr=2 - papyrus n= 5 confirmed-Egyptian= 5 with-inscription=0 total-inscr=0 Gathered in two passes: 1. 90 via `query={form_term}` + `department=Antiquities` (page 1 per category) 2. 20 supplemental via targeted queries (`stela hieroglyph`, `coffin hieroglyph`, `shabti hieroglyph`, `false door`, `book of the dead papyrus`) to surface known-inscribed Egyptian items the broad form-queries underranked. ### Top-level schema (detail endpoint) summary, identifier, admin, description, owners, title, institutions, lifecycle, multimedia, materials, techniques, name, inscription, categories, department, measurements, publications, type, apiURI, URI, object ### inscription[] — the headline structured field Per entry: { location, method, type, description (list-wrapped), transcription (list-wrapped) }. The transcription is Egyptian transliteration following MdC / Latin-Leiden conventions (Htp, di, nsw, etc.) — same form TLA emits. Means we can fuzzy-match Fitzwilliam inscriptions ↔ TLA lemma records by transliteration text without explicit cross-refs. ### lifecycle.collection.places (find spot, with coordinates!) 61/109 samples have ≥1 find-spot place; **50/109 have lat/lng coordinates as structured floats.** First museum source to expose georeferenced provenance directly. Place URIs (term-XXXXX) link into Fitzwilliam's controlled vocabulary (term-107870 is Egypt with 16,801 records). ### lifecycle.creation.periods (structured dynasties + chronology) 71/109 samples have structured periods. Values like "9th Dynasty", "Late Period", "Ptolemaic", "Coptic", "Third Intermediate Period". Multiple periods routinely co-present. Numeric chronology in `date.from.value` as signed ints (negative = BC). ### multimedia field (images at 3 resolutions) 48/109 samples have ≥1 image. Per-image: preview/large/original JPEG URLs at data.fitzmuseum.cam.ac.uk/imagestore/. ### Identifier types (multi-typed array) - `accession number`: E.6.1909 (Egyptian Antiquities: E.LOT.YEAR) - `priref`: 51004 (Axiell internal numeric) - `oldadmincategory`: legacy codes (SS, MS) - `Sketchfab`: 3D-model GUIDs (when present) - `uri`: full object URI ### Four-flag comparison (museum pattern now N=4) | metric | MET (#3) | BM (#4 via WD) | BKM (#5 via WD) | Fitz (#6 direct) | |---|---|---|---|---| | total surveyed | 77 | 57 (13 Egyptian) | 4 (4 Egyptian) | 109 (80 Egyptian) | | has inscription field | 0/77 | 0/57 | 0/4 | **6/109** ⭐ | | with transcription text | 0/77 | 0/57 | 0/4 | **6/109** ⭐ | | multi-inscription/object | n/a | n/a | n/a | yes (avg 4 on inscribed) ⭐ | | with TM cross-ref | 0/77 | 2/57 | 0/4 | 0/109 | | with TLA cross-ref | 0/77 | 1/57 | 0/4 | 0/109 | | with image | 77/77 | 48/57 | 4/4 | 48/109 | | with structured find spot | prose only | no | no | **61/109** ⭐ | | with find-spot coords | no | no | no | **50/109** ⭐ | | with structured period | period/dynasty fields | prose | prose | **71/109** ⭐ | | direct API access | yes (open) | no (Cloudflare) | no (Vercel) | yes (key required) | | coverage on direct path | 100% (27,972) | <0.1% (WD only) | <0.01% (WD only) | 100% (~16,801 Egyptian) | | WD coverage as fallback | irrelevant | 13/57 had WD | 4/4 had WD | 0 Egyptian via P8910 | ### Flag 1 — Object vs inscription modeling (REVISED) Previous (MET/BM/BKM): museums don't model inscriptions. **Revised (after Fitzwilliam): museums vary widely.** Some model inscriptions explicitly with transcription content (Fitzwilliam, and likely other Axiell/CIIM users); most don't. Our `object` ↔ `inscription` 1:N schema decision still holds — necessary for Fitzwilliam, harmless for others. Fitzwilliam validates the multi-inscriptions-per-object assumption directly (up to 7 inscriptions observed on one stela). ### Flag 2 — Transcription presence (REVISED) Previous: absent across MET/BM/BKM. **Revised: Fitzwilliam is on the TLA/TM end of the spectrum — museum-side records carry transliteration directly.** Fitzwilliam becomes a *primary source* for inscription text in Phase 2, not just a join target. ### Flag 3 — Cross-ref gap (CONFIRMED, but transcriptions enable a workaround) Direct API exposes no TM/PM/JdE. Wikidata covers 1,711 Fitzwilliam items via P8910 but 0 with Egyptian signals. **Workaround unique to Fitzwilliam**: because both Fitzwilliam and TLA emit transliterations in compatible conventions, we can fuzzy-match Fitzwilliam inscriptions to TLA lemma records by transliteration text. Build cross-refs ourselves without waiting on either side to expose them. Also: Fitzwilliam's `publications` field references academic catalogues per object. Those catalogues often cite TM and Porter-Moss numbers — mining publication metadata is another route to discover cross-refs. ### Flag 4 — Image licensing Metadata is **CC-BY-NC-ND 4.0** (per API ToS). Per-image rights vary. 48/109 samples have ≥1 image. ### Operational gotchas - **Signup endpoint**: POST `{name, email, password}` (OpenAPI `required` is empty but all three must be sent). JWT is Laravel Sanctum: `"{id}|{token}"` — works as standard Bearer. - **`category` and `place` params reject all values tested** (HTTP 400). Only `query` (free-text) and `department` work. - **`query` ranks by search relevance, not Egyptian-ness** — `query=stela` returns mostly amulets first. Use targeted queries like `query=stela hieroglyph` to surface inscription-rich items. - **Rate limiting**: HTTP 429 after ~70 rapid detail fetches. Use 0.8s sleep + 30s exponential backoff on 429. - **/iiif/{object} endpoint rejects all ID formats tried** — multimedia URLs work without IIIF, so we punt on this. ### Open questions - IIIF endpoint format — would enable OpenSeadragon viewers. - `publications` cross-refs to academic catalogues — likely route to Porter-Moss / Bibliography of Egyptian Topography references. - The `terminology` endpoint's controlled vocabulary (term-XXXXX) is worth mining for normalised place/period hierarchies — useful for standardizing find spots across sources. - High-inscription-rate object types (stelae, coffins, BoD papyri): what proportion of Fitzwilliam's Egyptian collection has populated transcription? A dedicated probe could move the museum-pattern needle.

Fitzwilliam Museum — Department of Antiquities (Egyptian)

Sample objects

stelae

Panathenaic amphora

Ptah-Sokar-Osiris box

Ptah-Sokar-Osiris box

Rod fragment

Rod fragment

adze

adze

adze

alabastron

alabastron

alabastron