Linked Open Art Data Web App (AI) — Must-have Data Sources

Purpose: a curated shortlist of high-leverage open / machine-readable sources to power a linked open art data app (entity resolution, enrichment, search, recommendations).

Recommended pattern (why this stack works)

• Use a backbone entity graph for identity + disambiguation.
• Use authority vocabularies for normalization (names, places, materials, techniques, concepts).
• Add institutional collection APIs for canonical object records, images, provenance, exhibitions.
• Normalize all ingested data into a common internal model (recommend: [[Linked Art]] style modeling).

Top 10 prioritized sources

Core art sources

1. [[Wikidata]]

• Best for: global entity backbone (artists, artworks, movements, places, institutions)
• Strength: broad linked-data connectivity; identifiers to many external catalogs
• Use in product: entity graph + join keys + disambiguation + “sameAs” expansion

1. [[Europeana]]

• Best for: aggregation/discovery across European cultural heritage
• Strength: Search/Record/Entity APIs for discovery + entity matching
• Use in product: broad coverage, cross-institution linking, discovery funnels

1. [[Rijksmuseum]] (Rijksmuseum Data)

• Best for: high-quality museum metadata, Europeana-aligned structures
• Strength: open museum dataset + APIs/downloads; strong object/creator metadata
• Use in product: deep collection depth, high-trust records for training/eval

1. [[The Met Collection API]] (Metropolitan Museum of Art)

• Best for: broad Open Access artworks + images
• Strength: popular and well-documented; good for rapid prototyping
• Use in product: canonical object pages, image-based features, recommendations

1. [[Art Institute of Chicago API]]

• Best for: unified public API spanning collection + related content
• Strength: consistent API surface; strong for integrating “content around art”
• Use in product: object + interpretive content; richer on-site experiences

High-value linked-data & standards

1. [[Getty Vocabularies]]

• Best for: authority control (names/places/materials/techniques/concepts)
• Strength: normalization + disambiguation; improves entity linking quality
• Use in product: canonical labels, multilingual expansion, semantic search facets

1. [[Linked Art]] (data model ecosystem)

• Best for: interoperability standard for art data publishing/integration
• Strength: consistent graph model; reduces bespoke mapping per museum
• Use in product: internal canonical schema + export format

1. [[Smithsonian American Art Museum LOD]] (SAAM)

• Best for: stable URIs for artists/objects; strong American art coverage
• Strength: linked open data with resolvable identifiers
• Use in product: graph linking, authority-style references, provenance context

1. [[Tate collection data]]

• Best for: open collection metadata around artists and artworks
• Strength: good coverage; useful for cross-museum linking
• Use in product: UK/modern art depth; complementary to Europeana + Wikidata

1. [[American Art Collaborative]] (and similar museum consortia LOD)

• Best for: cross-museum linking beyond a single institution
• Strength: multi-institution graph context; increases match rates + coverage
• Use in product: broader graph + better recommendations across collections

Practical build order (implementation sequence)

1) [[Wikidata]] for the backbone entity graph + disambiguation.

2) [[Getty Vocabularies]] for controlled terms / normalization.

3) [[Europeana]] for broad aggregation + entity matching.

4) Add “depth” via: [[Rijksmuseum]], [[The Met Collection API]], [[Art Institute of Chicago API]], [[Smithsonian American Art Museum LOD]], [[Tate collection data]].

5) Adopt [[Linked Art]] as the canonical internal model and map sources into it.

Notes for AI-powered features (what to extract/standardize)

• Identity: stable IDs, sameAs links, external identifiers
• Strings to normalize: person names, place names, titles, materials, techniques
• Time: production dates (with uncertainty), life dates, period styles
• Relationships: creator ↔ work, work ↔ movement, work ↔ place, work ↔ institution
• Assets: image URLs, IIIF manifests (when available), rights statements

Follow-ups / next steps

• Create individual notes per source with: endpoints, rate limits, licensing, identifier strategy, mapping notes to [[Linked Art]].
• Define internal entity types (Artist, Artwork, Place, Movement, Institution, Material, Technique) and linking rules.