Voting closed 9 May, 2013. 164 Liked
Title: Linked Data Data Modeling: Converting VRA Core 4 into Linked Data
Short Description: This project attempted to bridge the gap between the founding principles of Linked Data and published RDF datasets. In doing so, it is hoped that Linked data can be more effectively used and implemented by Galleries, Libraries, Archives and Museums. The focus of this study was developing a Linked Data data model that incorporates popular domain specific vocabularies (primarily Schema.org and FOAF). The VRA 4 restricted XML schema was converted into a new Linked Data data model using Protégé, then an XSLT stylesheet was used to convert an existing VRA 4 dataset into RDF. The detailed description of the study as well as the tools developed (including the Ontology and the XSLT stylesheet) can be downloaded from http://purl.org/jmixter/thesis
Long Description: Linked Data has become an increasingly important and valuable way for sharing data across the Internet. It is the basis for the Semantic Web and allows organizations to not only easily share data, but also connect data with other related data. Visual Resource Association (VRA) Core 4 is an XML schema-based data model for cataloging cultural objects and visual resources. Using the existing VRA Core 4 restricted XML schema, a new data model was developed that took advantage of popular domain specific vocabularies. Using popular vocabularies such as Schema.org, helps ensure that data will be interoperable with other data and can potentially help improve visibility on the Internet. Using the data model as a reference, an ontology was developed using Protégé ontology editor. It illustrated how popular domain specific vocabularies can be combined with the existing VRA data model to create a new semantically-rich model that still retains the specificity and detail of the original XML restricted schema. In addition to developing a new VRA data model, an XSLT stylesheet was created that demonstrated how existing XML based records could be converted into RDF data. The stylesheet was used to successfully convert a 4,150 record collection from the University of Notre Dame into RDF triples. The XSLT templates used in the stylesheet were able to not only convert the existing XML elements/attributes into RDF classes/properties but also convert the existing controlled vocabulary terms into functioning http URIs representing concepts. The study successfully demonstrated that existing data models can be enhanced to incorporate Linked Data and that existing datasets of implementation-specific XML records can be converted into RDF triples with properties defined by popular RDF vocabularies using an XSLT stylesheet.
Developed Tools: Everything that was developed for this project is free to use and download from the above website. These tools include
- A proposed VRA Ontology (available in both OWL format as well as HTML)
- An XSLT stylesheet that can be used to convert existing VRA/XML data
- Published paper that explains how the study was developed and conducted and explains, in detail, the results