GloPadRDF

Attempt to move GloPad into RDF and then into some visualization tool.

Project for Technology Innovation Week - Jan 6 2015 Brian Caruso, Shinwoo Kim, James Reidy.

Converting the GloPAD data to RDF

The GloPAD data was provided as a dump of a postgres database. This was loaded into a postgres 9.3 database and then mapped to RDF using D2RQ.

To generate mappings and deploy to tomcat

$ cd d2rq
$ ./glopad-table-analysis.pl yourdbuser dbpassword
# result is placed in webapp/WEB-INF/mapping.ttl
$ ant war
$ cp d2rq.war /lib-dev/tomcat/webapps

To dump GloPAD as N3 RDF

$ cd d2rq
$ ./glopad-table-analysis.pl bdc34 xyz
# result is placed in webapp/WEB-INF/mapping.ttl
$ ./dump-rdf -b http://bdc34-dev.library.cornell.edu:8080/GloPAD/ webapp/WEB-INF/mapping.ttl > glopad-dump.n3

To generate an ontology for the RDF

The glopad-dump.n3 doesn’t have classes and properties defined. Pellet can generate them. See: http://clarkparsia.com/pellet/icv/

$ cd ~/pellet
$ ./pellet.sh lint -v -only RDF -f /lib-dev/vitro-glopad/lopad-ont.rdf ~/glopadRdf/glopad-dump.n3 > /lib-dev/vitro-glopad/glopadLint.out

Some tasks that have not been done

The mapping created during the week in 2015-01-20 is a rough prototype and not a complete working example.

  1. An ontology has not been created. A quick one was created with Pellet.
  2. The IRIs used do not work well with RDF/XML tools such as
  3. The classes and properties are not associated with any popular ontology. Since we are working with relations that come up in many contexts we should be able to find existing ontologies to use. Some to consider are DBPedia, FOAF and BIBFRAME.
  4. The images from the GloPAD system are not represented well.
  5. The GloPAD alt names and multilingual names are not mapped.
  6. The GloPAD dates are not mapped, there is a first attempt to map them in datesMapping.ttl.

Visualization of GlopadRDF

See contents of lodlive directory.

Why convert GloPAD to RDF?

GloPAD is offers a wealth of relations between digital objects, people involved in the performing arts, plays, performances and theaters. If this collection of information is valuable, it will be necessary to maintain it to continue to provide access to it. Converting the data in GloPAD to RDF offers a way to performance this maintenance with advantages over alternatives.

  1. System agnostic data format avoids lock in
  2. Possible to use well known properties such as from DBPedia or BIBFRAME
  3. Many tools used by CULIT work with RDF, drupal, Ruby on Rails, Vivo, Fedora

Ideas for further work

  1. Map the GloPAD data better. Get the dates, multilingual labels, URLs for images etc.
  2. Once the data is mapped to RDF, transform as many properties and classes as possible to commonly used ontologies with SPARQL constructs, or SWRL.
  3. Associated individuals in the GloPAD RDF with individuals in DBPedia, LOC, or other LOD sources.
  4. Load the data into Vitro using rdf/abox/firsttime and rdf/tbox/firsttime.
  5. See if editing and display in Vitro are acceptable and create custom forms and templates where they are not.
  6. Asses other RDF tools for use with GloPAD such as Drupal or ActiveRDF+RoR.