Based on the Linked Art documentation, here’s a synthesis of the key patterns used throughout the model:
Core Foundational Patterns
- Activity/Event Pattern
Linked Art uses intermediate entities representing activities to connect other entities, rather than direct relationships. For instance, instead of directly linking an artwork to an artist, there’s a creation activity carried out by the artist that can have additional properties like time, place, techniques, and other objects . This creates a “What, Who, Where, When, Why and How” framework.
- Classification Pattern
The model extensively uses the classified_as property to type entities using controlled vocabularies (particularly the Getty AAT). This allows flexible categorization without creating specific relationships for every case. Concepts can have both broader terms and classifications, with broader terms encompassing the concept directly while classifications categorize it .
- Core Properties Pattern
Every entity has essential properties for usability:
• Unique identifier (id)
• Type classification (type)
• Human-readable label (_label)
• Names and identifiers (identified_by)
- Statement Pattern
When data doesn’t support full specificity or is best expressed in human-readable form, the pattern uses the LinguisticObject class to record descriptive text that can be referenced by any other entity .
- TimeSpan Pattern
The minimal timespan model uses begin_of_the_begin and end_of_the_end properties with fully qualified ISO 8601 timestamps, plus human-readable Names for display .
- Expansion Pattern
Primary information uses simple patterns while more complex data expands upon it without interference - for example, basic Production uses carried_out_by for the Artist, while previous attributions use AttributeAssignment to expand upon this .
These patterns work together to create a flexible, event-based system for describing cultural heritage that balances machine-readability with practical usability.
Based on the Linked Art model and real-world implementations, here’s how to best leverage it for websites and web apps with optimal UI/UX:
Technical Implementation Strategy
- API-First Architecture
• Use the Linked Art API endpoints as your data layer
• Implement JSON-LD consumption on the frontend (it’s JSON, so JavaScript-friendly)
• Cache frequently accessed relationships to reduce API calls
• Use the search API with HAL links for discoverability
- Progressive Enhancement Approach
The exhibition browser exemplar demonstrates creating sustainable presentations with minimal resources - ideal for small institutions or projects with limited maintenance capacity . Start simple and layer complexity.
- Data Transformation Layer
• Transform Linked Art’s event-based model into user-friendly views
• Pre-compute common queries (artist’s works, object provenance chains)
• Use the classification system to drive filtering and navigation
UI/UX Best Practices
Display Patterns for Core Entities:
Objects/Artworks:
• Lead with high-quality images and essential metadata (title, artist, date)
• Use expandable sections for production details, materials, dimensions
• Show provenance as a visual timeline rather than dense text
• Link related works through “Explore Similar” sections
Artist/Agent Pages:
• Biography with timeline of key activities (productions, exhibitions)
• Grid of associated works
• Map showing places of activity/residence
• Network visualization of relationships with other artists
Exhibitions:
Museum apps can use linked data to provide richer contextualized information - visitors scan QR codes to access biographical information, artwork history, and connections to related works . Create immersive exhibition experiences both in-person and online.
Search & Discovery:
• Faceted search using classifications (medium, time period, style)
• Auto-suggest powered by identifiers and alternate names
• “Related to this” sections powered by the activity connections
• Visual browsing through image grids before detail views
Specific UX Recommendations
- Handle Complexity Gracefully
The Art Tracks project defined four complexity levels for provenance, treating simpler levels as subsets so records can be displayed at appropriate granularity based on available data . Apply this concept broadly:
• Show basic info by default
• Offer “See full details” for researchers
• Provide both human-readable and technical views
- Leverage Relationships for Navigation
The event-based model creates rich connections - turn these into intuitive navigation:
• “Who else exhibited here?”
• “What else was made in this technique?”
• “Where else did this object travel?”
- Multi-Modal Access
• Timeline views for chronological data
• Map views for geographic data
• Network graphs for relationships
• Traditional list/grid for browsing
- Statement Pattern for User Experience
When detailed data isn’t available, Linked Art uses descriptive text. In UI:
• Display these prominently with attribution
• Use them as fallbacks when structured data is missing
• Allow toggling between “what we know” and “what’s documented”
Performance Considerations
• Pagination: Use the API’s built-in pagination for large result sets
• Lazy Loading: Load detailed relationships on-demand
• Image Optimization: Leverage IIIF for responsive images if available
• Caching Strategy: Cache entity pages but invalidate when relationships change
Accessibility & Internationalization
• The _label property is perfect for screen readers
• Multiple Name objects support multilingual content
• Classifications link to Getty AAT which has translations
• Provide text alternatives for all visualizations
Would you like me to create a prototype interface demonstrating any of these patterns?