Skip to end of metadata
Go to start of metadata

LD4L2 grant proposal, Here is one set of ideas we could propose contributing:

  • Create a linked-film web site based on the Harvard Film Archive. Michele Durocher at Harvard is exploring this now.
    • Select the base ontology, db-pedia, the movie ontology, other
    • Create mapping rules to map HOLLIS records for film to the ontology
    • Deploy as linked data in the Harvard triple store
    • Create a user interface/ visualization of the data that works for film researchers
  • Visual material metadata as linked data
    • Create a converter for MODS (or ARTstor VRA) --> LD4L-BIBFRAME
    • Export all Harvard’s 2M visual information records as LD4L BIBFRAME RDF linked data
    • Enrich with entity resolution, taking advantage of  ARTstor Shared Shelf’s existing authority linkage for authority links (Getty, etc.) in visual data
    • Expose as linked data at Harvard with a SPARQL endpoint
    • Export the RDF to the Cornell/Stanford hub.
  • Archival finding aid metadata as linked data
    • Create a converter for MODS --> LD4L-BIBFRAME
    • Export all Harvard’s 4M finding aid components as LD4L BIBFRAME RDF linked data
    • Enrich with entity resolution, taking advantage of a TBD EAC-CPF authority
    • Expose as linked data at Harvard with a SPARQL endpoint
    • Export the RDF to the Cornell/Stanford hub.
  • GIS metadata as linked data
    • Create a converter for FGDC --> LD4L-BIBFRAME
    • Export metadata for all Harvard’s10,000 fGIS data sets as LD4L BIBFRAME RDF linked data
    • Enrich with entity resolution, taking advantage of Getty TGN, LC Names, and DB Pedia place records
    • Expose as linked data at Harvard with a SPARQL endpoint
    • Export the RDF to the Cornell/Stanford hub.
    • Leverage GeoHOLLIS spatial location data associated with catalog records to link to spatial resources on the web
  • :Linked data discovery API and widget
    • Develop  back end inferencing middle ware, to gather relevant links for a given record and expose them as a REST API with a limited graph
    • Develop a javascript widget that uses the API to create a textual and graphical visualization of most relevant links for a catalog record
  • No labels