Linked Data Elements at COL

Interacting with example data in a linked data environment with emphasis on the SPARQL endpoint

Press key to advance.

Having issues seeing the presentation? Read the disclaimer

Linked Data Demo

Initial Primary Data Sources

  • Age Models ( (derived from Janus))
  • Image Strips (LDEO (derived from Janus))
  • Janus Drill Down LD demo app (hosted at COL (derived from Janus))
  • Log Files (hosted at LDEO)

Other graphs in the pipeline

  • Foram data (CHRONOS hosted Foram data)
  • Supporting vocabularies
  • Timescale data (CHRONOS hosted)
  • SIOC from web sites
  • Reports
Linked Data Demo

Data Flow

Approaches for Demo Phase

  • Web apps generate graphs for pages and also "bulk" graphs
  • These graphs are "scrapped" from the app or in the case of "bulk" graphs simply downloaded
  • Then loaded into the SPARQL server using application tools

Approaches for Production

  • Data pull (harvesting)
  • Data push (applications push changes (diffs) up)

SPARQL interface

SPARQL end point SPARQL is a topic unto itself!

SELECT ?uri 
   ?uri skos:related <> .
PREFIX chronos: <>
SELECT ?uri ?long ?lat
     ?uri skos:related <> .
     ?uri geo:long ?long .
     ?uri geo:lat ?lat .
Linked Data Demo

Content Negotiation

Approaches for Testing

  • The URI represents the resource, but not necessarily the representation
  • The representation can be negotiated
  • Typically at least XHTML and RDF

Examples of negotiating URI's (Accept:Vary)

  • curl --header "Accept:text/html" -o ~/Desktop/test.html
  • curl --header "Accept:application/rdf+xml" -o ~/Desktop/test.rdf
  • curl --header "Accept:image/jpeg" -o ~/Desktop/test.jpeg
Linked Data Demo


A few items to consider

  • Linked Data != Semantics (but it's a good starting point)
  • We didn't review RDFa and other important aspects
  • Work on relations built on concepts and resources (
  • Searches can be "federated" and results point to N authoritative resolvers
  • The interface/search can be exposed via web arch and exposed by many clients or means
  • Full web arch so we can cache, proxy, 303 etc.
  • What goes in the graph and what doesn't?
  • Interfacing to other ID mechanisms like DOI's
Linked Data Demo