Presented by Jeffrey Mixter at the Annual Conference of the Visual Resources Association, March 12-15, 2014 in Milwaukee, Wisconsin.
Session 11, Brave New World Cataloging: Using RDF and Linked Open Data for the Semantic Web
ORGANIZER: Sheryl Frisch, California Polytechnic State University
MODERATOR: Trish Rose-Sandler, Center for Biodiversity Informatics, Missouri Botanical Garden
PRESENTERS:
• Trish Rose-Sandler, Center for Biodiversity Informatics, Missouri Botanical Garden
• Jeffrey Mixter, Kent State University Research Support
• Georgina Goodlander, Smithsonian American Art Museum
• Patricia Harpring, Getty Vocabulary Program, Getty Research Institute
RDF (Resource Description Format) and LOD (Linked Open Data) are two key components in the ongoing development of the Semantic Web (the structured linking of web-based information to enable users anywhere to find, share, and combine information more easily). Although we are used to working in information silos much of the time, the Semantic Web can allow data to be discovered, shared and reused across application, enterprise, and community boundaries. The speakers will demonstrate how our existing data (from both VR collections and museums) can be transformed to the RDF format; how the effort can be shared in a community; and how LOD will affect and expand the tools we use daily to provide controlled vocabulary terms.
Using Grammatical Signals Suitable to Patterns of Idea Development
VRA 2014 Brave New World Cataloging, Mixter
1. VRA Core as Linked
Data
03-15-2014
Jeff Mixter
Research Support Specialist
OCLC Research
mixterj@oclc.org
2. • This project evolved out of a master’s thesis project published
as partial completion of an MLIS degree from Kent State
University
• The emergence of Schema.org as a general Linked Data
vocabulary and the publication of WorldCat.org data as Linked
Data served as a catalyst for the work
• CIDOC CRM and the Europeana Data Model were examples of
how visual resource data models could be represented using
Linked Data
Background
3. • Collaborative effort between Google, Yahoo, Bing and Yandex
• 2011
• Serves as a general Linked Data vocabulary
• 15% of Web pages use Schema.org markup
• Data using Schema.org is understood and indexed by search
engines as structured data
• Using Schema.org has the possibility to help improve SEO
• SEO in context!
• SEO != Google Ranking
• Search engine understanding of your structured data
Schema.org
4. • Structured data drives the
creation of Google
Knowledge Cards
• These “entity cards” are
derived from structured
data that Google has
harvested from the
Semantic Web
Structured Data
6. • Both of these data models have produced Linked Data
representations
• Both use a specialized vocabulary
• The models were reviewed and consulted during the
modeling/mapping process
CIDOC CRM and Europeana
7. • The initial research project sought to map the VRA Core 4 model
into Linked Data
• Use Schema.org as the baseline vocabulary
• Helps improve interoperability
• Follow the Linked Data principles outlined and described by Tim
Berners-Lee
• Do not reinvent the wheel
• Gov 2.0 Expo 2010
• The project successfully modeled the VRA Core 4 data model as
Linked Data and demonstrated this by converting an existing
VRA Core 4 compliant dataset (XML) into Linked Data
Initial Project
8. • Linked Data model was mapped from the VRA Core 4 Restricted
Schema
• The model included:
• 35 Schema.org terms
• 1 DC Terms term
• 2 VoID terms
• 127 Custom VRA Core terms
• Of the 127 custom terms, 88 (66%) were positioned as sub-terms of
Schema.org terms
• Overall 74% of the VRA Core terms were directly or indirectly
mapped to Schema.org
Project Results
9. • After the prototype and results were published, there was
interest in using the model
• The VRA Core Oversight Committee subsequently formed a
Task Force to design and publish as VRA Linked Data data
model
• Current members:
• Esme Crowles
• Trish Rose-Sandler
• Johanna Bauman
• Rebecca Guenther
• Jeff Mixter
Aftermath
10. • Adapt the prototype model to create an official VRA Core data
model
• Using a VRA namespace!
• In line with the principles of Linked Data, there will be an
emphasis on mapping the VRA Core model to other Linked
Data vocabularies
• Schema.org
• BIBFRAME
• Ect.
Current Project
11. • Prototype of the ontology:
• http://www.essepuntato.it/lode/https://s3.amazonaws.com/VRA/
Ontology/VRA_OntologyRevised.owl
• Converted a data sample for review
• The model is just a prototype and we welcome comments,
questions, criticism, etc.
Progress
12. • Getty vocabularies are being published as Linked Data
• This is very important to the micro-domain of visual/cultural
heritage resources
• In particular it is important to VRA since it is a standard authority
recommended in the VRA Core 4 Schema
• Collaboration and integration with other Linked Data models
• The Semantic Web relies on interconnected datasets/models
• Data silos are bad
Future Potential