Linked Open Data for the Cultural Sector

1,505 views

Published on

About how

Published in: Technology, Education
0 Comments
4 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,505
On SlideShare
0
From Embeds
0
Number of Embeds
53
Actions
Shares
0
Downloads
4
Comments
0
Likes
4
Embeds 0
No embeds

No notes for slide

Linked Open Data for the Cultural Sector

  1. 1. Cultural Linked Open Data 2014-02-06 Lars Marius Garshol, larsga@bouvet.no, http://twitter.com/larsga 1
  2. 2. The importance of data • Most web sites are data-driven – if you have the data, you can add functionality – if you don’t have the data, you’re stuck • Example: Google Maps – imagine you have the application, the server farm, the scaling and monitoring, etc – but you don’t have the actual map data – not only are you stuck, but creating the data is much harder than making the service 2
  3. 3. 3
  4. 4. 4 Research project by SINTEF and Computas
  5. 5. Data sources Research project by SINTEF and Computas 5
  6. 6. Must be at meeting at 1345. Three transport alternatives. 6 Research project by SINTEF and Computas
  7. 7. Data is raw material for building services! 7
  8. 8. Possible users of cultural data • Any kind of web store – publishers – streaming services – ... • Travel businesses – public sector, hotels, tour organizers, event organizers, ... • Media – newspapers, broadcasting, ... • Lots of public sector uses – education, ... • Many things none of us can’t imagine now 8
  9. 9. 9
  10. 10. Only linked data is usable NRK/Skole Cappelen Damm 10
  11. 11. Linked Open Data • Movement to publish open data online – in machine-readable form – linked to other data sets • Based on some key technologies – URLs for identifiers – RDF for data • Gaining a lot of traction in the cultural sector – – – – 11 BBC Europeana Smithsonian Institution ...
  12. 12. The technology • Provides simple data representation – – – – graph model (RDF) has ready-made formats (XML, text, JSON, ...) standard query language (SPARQL) lots of RDF databases available • Allows anyone to refer to anything – a museum can say explicitly that one object in their collection has a specific relation to an object in another collection – liberation from the ID scheme confusion • Can reuse terminology from other authorities – can also easily extend that terminology 12
  13. 13. 13 http://lod-cloud.net/
  14. 14. 14
  15. 15. http://dbpedia.org/resource/Knut_Faldbakken • Globally unique – across all systems and organizations • Distributed – if you have a domain, you can make URIs • Self-documenting – just follow the link to find documentation • Can be used anywhere – anyone can point at anything
  16. 16. Today • • • • Flat, unlinked data No navigation No connections Poor characterization – doesn’t say what it is 16
  17. 17. Europeana Data Model As linked data edm:ProvidedCHO nv:Photograph rdf:type dc:title dc:date “Bergliot Ibsen” 1903 dc:subject foaf:Person rdfs:label “Bergliot Ibsen” dbp:died 1953-02-02 dbp:born 1869-06-10 nv:provider http://dbpedia.org/resource/Bergliot_Ibsen rdfs:label grs:point 17 “Aulestad” 61.2173 10.265952 http://dbpedia.org/resource/Aulestad
  18. 18. Choice of tools Modelling pellet Reasoners Redland RDF Libraries APIs Triple stores
  19. 19. Great, but how can we actually link the data? 19
  20. 20. 20
  21. 21. “Do they have Knut Faldbakken in here?” 21 http://data.deichman.no/sparql
  22. 22. Yes, but not connected to anything ... ...can we do anything about that? 22
  23. 23. Record linkage to the rescue • Active research field – dating back to the 1940s • Can connect data without common IDs – measure similarity instead • Tools exist, with – – – – value cleaning statistical analysis sophisticated comparators fast search backends • One example is Duke – http://code.google.com/p/duke/ – Java and open source 23
  24. 24. Connect to DBpedia http://data.deichman.no/...dbakken_Knut_1941- http://dbpedia.org/resource/Knut_Faldbakken NAME: LIFESPAN: NATIONALITY: n NAME: BIRTHDATE: Faldbakken, Knut 1941- Knut Faldbakken 1941-08-31 Complete recipe here 24 http://code.google.com/p/duke/wiki/DeichmanLink
  25. 25. Training with genetic algorithm 25 http://www.garshol.priv.no/blog/262.html
  26. 26. Conclusion • Linked Open Data has tremendous potential – vastly easier reuse of data – hugely empowering for consumers – also opens new possibilities for data owners • Growing use in cultural sector – both internationally and in Norway • To learn more – http://www.slideshare.net/larsga/linked-opendata-14964163 – http://data.norge.no/veiledning – http://linkeddatabook.com/editions/1.0/ 26
  27. 27. Hafslund SESAM 27

×