Project Ideas

If you are interested in collaborating with DBpedia Spotlight's page describes some project ideas on which we'd like to work with you

• Project Ideas for the ISSLOD 2011
  • Description: http://pablomendes.wordpress.com/2011/09/13/lod2-indian-summer-school-project-ideas/
  • Project sizes: 3-5 days, part-time
• Project Ideas that we suggested for the Google Summer of Code 2012:
  • Description: GSoC2012-Project-Ideas
  • Project sizes: 3 months
• Project Ideas for the AISLOD 2012
  • Description: (see below)
  • Project sizes: 2 days, part-time

Project Ideas for the AISLOD 2012

facetnator

Given a list of textual HTML elements (e.g. news articles), automatically organize them by topic, by person mentioned, etc. -- providing a menu of facets to browse through them

• send text to DBpedia Spotlight
• collect the annotations for each text item
• optionally, enhance annotations by fetching more information (e.g. categories using SPARQL query to DBpedia endpoint -- property is probably skos:subject)
• create menus to filter content
• as users click on menus, filter the content appropriately

An example facet browser is here: http://www.simile-widgets.org/exhibit/

Suggestion: use jQuery (there is a jQuery plugin for DBpedia Spotlight) http://dbpedia-spotlight.github.com/demo/dbpedia-spotlight-0.3.js

rel8

Given one URI, provides a list of other related URIs based on DBpedia Spotlight's definition of semantic relatedness (similar contexts). Let's say you are interested in things related to Biofuel: http://spotlight.dbpedia.org/related/?uri=Berlin

This has been computed by looking at DBpedia Spotlight's model of a resource based on the words that appear around them in Wikipedia. You could enhance this by using SPARQL queries to obtain more info and refine this definition.

• type: only allow URIs that are similar AND belong to the same class in the DBpedia ontology
• links to other entities: use DBpedia wikiPageLinks relation to compute how many DBpedia Resources link to both resources and count those as a measure of similarity
• categories: use the skos:subject property to compute how many categories are shared by both resources

Bits and pieces that are already done

Testing the deployed demo

 curl -H "application/json" "http://spotlight.dbpedia.org/related/?uri=Berlin"

Getting the code

Starting the server

  mvn scala:run -DmainClass="org.dbpedia.spotlight.web.rest.RelatedResources"

Using the server

  curl -H "application/json" "http://localhost:2222/related/?uri=Berlin"

Using the sparql endpoint with GSoC projects

The command below uses cURL to execute a SPARQL query that retrieves all GSoC projects tagged with the string "css".

curl http://spotlight.dbpedia.org/sparql/ -d "query=select * where { ?s  <http://spotlight.dbpedia.org/gsoc/vocab#taggedString> \"css\"@en } limit 5"