Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Linked Logainm: Enhancing Library Metadata using Linked Data of Irish Place Names

633 views

Published on

Presentation at the First Workshop on Linking and Contextualizing Publications and Datasets

Published in: Technology, Education
  • Be the first to comment

  • Be the first to like this

Linked Logainm: Enhancing Library Metadata using Linked Data of Irish Place Names

  1. 1. Digital Enterprise Research Institute www.deri.ie Enabling networked knowledge Linked Logainm: Enhancing Library Metadata using Linked Data of Irish Place Names Nuno Lopes Rebecca Grant Brian Ó Raghallaigh Eoghan Ó Carragáin Sandra Collins Stefan Decker September 26, 2013
  2. 2. logainm.ie The authority list of Irish place names, validated by the Placenames Branch. Delivering a more detailed level than in DBpedia, Geonames. Unique source of Irish language place names 1 / 13
  3. 3. logainm.ie The authority list of Irish place names, validated by the Placenames Branch. Delivering a more detailed level than in DBpedia, Geonames. Unique source of Irish language place names But.. not easily accessible automatically 1 / 13
  4. 4. The NLI Longfield Map Collection The Longfield Maps are a set of 1,570 surveys carried out in Ireland between 1770 and 1840. Currently catalogued in MarcXML Integrating Logainm data into their workflow: for enabling searching for place names in Irish using Linked Data 2 / 13
  5. 5. Longfield Map example 3 / 13
  6. 6. Longfield Map example MARC/XML <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land tenure</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Rathdown (Barony)</marc:subfield> </marc:datafield> <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land use surveys</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Wicklow (County)</marc:subfield> </marc:datafield> 3 / 13
  7. 7. Approach for creating the dataset 1 Translate Logainm database dump into RDF 2 Determine links to other datasets based on: Place names Type Geographical coordinates Hierarchy of places 3 Evaluation of generated links 4 Library catalogue enhancement 4 / 13
  8. 8. Overview of GLD Providers: DBpedia Exported from Wikipedia LinkedGeoData Exported from OpenStreetMap GeoNames 5 / 13
  9. 9. Overview of GLD Providers: DBpedia Exported from Wikipedia LinkedGeoData Exported from OpenStreetMap GeoNames GeoLinkedData Ordnance Survey 5 / 13
  10. 10. Overview of GLD Providers: DBpedia Exported from Wikipedia LinkedGeoData Exported from OpenStreetMap GeoNames GeoLinkedData Ordnance Survey Vocabularies: W3C Geo SpatialThing NeoGeo Feature vs Geometry Spatial Relations (is_part_of) Most providers define their own 5 / 13
  11. 11. 1. Converting Logainm dump to RDF SPA QL M L X D F R ∼ 1.3M triples Data provided in XML 6 / 13
  12. 12. 1. Converting Logainm dump to RDF SPA QL M L X D F R ∼ 1.3M triples Data provided in XML Translated to RDF using XSPARQL 6 / 13
  13. 13. 1. Converting Logainm dump to RDF SPA QL M L X D F R ∼ 1.3M triples Data provided in XML Translated to RDF using XSPARQL Exposed using Openlink Virtuoso 6 / 13
  14. 14. Linked Logainm http://lod-cloud.net/ Government Media User-generated Publications Life sciences Cross-domain GeoLogainm OCLC FAST 7 / 13
  15. 15. Linked Logainm http://lod-cloud.net/ Government Media User-generated Publications Life sciences Cross-domain GeoLogainm OCLC FAST 7 / 13
  16. 16. Linked Logainm http://lod-cloud.net/ Government Media User-generated Publications Life sciences Cross-domain GeoLogainm OCLC FAST 7 / 13
  17. 17. 2. Place name matching using Silk 1 Place Name Island, Cavan: 2641 "Place"s in DBpedia Airport, Dublin: 7828 8 / 13
  18. 18. 2. Place name matching using Silk 1 Place Name Island, Cavan: 2641 "Place"s in DBpedia Airport, Dublin: 7828 2 Geographical Location ∼50% of place names in logainm contain geographical information 8 / 13
  19. 19. 2. Place name matching using Silk 1 Place Name Island, Cavan: 2641 "Place"s in DBpedia Airport, Dublin: 7828 2 Geographical Location ∼50% of place names in logainm contain geographical information 3 Name of the county / parent place name 8 / 13
  20. 20. 2. Place name matching using Silk 1 Place Name Island, Cavan: 2641 "Place"s in DBpedia Airport, Dublin: 7828 2 Geographical Location ∼50% of place names in logainm contain geographical information 3 Name of the county / parent place name 4 Mapping of types from Logainm to types in other datasets logainm.ie DBpedia LinkedGeoData Geonames townland Populated Place Locality LCTY, PPLF 8 / 13
  21. 21. 3. Silk results Entities IE # Links % Links DBpedia1 10,715 1,552 14.5 LinkedGeoData2 36,237 6,611 18 GeoNames3 23,102 8,229 35.5 1 Entities of type “Place” or “Feature” 2 Entities of type “Node” 3 No hierarchy info 4 Including internal & Freebase links 9 / 13
  22. 22. 3. Silk results Entities IE # Links % Links DBpedia1 10,715 1,552 14.5 LinkedGeoData2 36,237 6,611 18 GeoNames3 23,102 8,229 35.5 Links in other datasets Entities # Links % Links DBpedia 873,643 653,7074 74.84 LinkedGeoData 6,251,067 462,098 7,4 1 Entities of type “Place” or “Feature” 2 Entities of type “Node” 3 No hierarchy info 4 Including internal & Freebase links 9 / 13
  23. 23. Evaluation Results Links Checked Correct DBpedia 1,552 1,552 (100%) 98% LinkedGeoData 6,611 500 (7.5%) 96% GeoNames 8,229 500 (6%) 99% Same place names can be “towns”, “population centre”, and “townland” in logainm.ie. DBpedia contains only one entry: Adrigole (population centre) and Adrigole (townland) http://dbpedia.org/resource/Adrigole Similar for LinkedGeoData 10 / 13
  24. 24. Longfield Map example (Updated) 11 / 13
  25. 25. Longfield Map example (Updated) <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land tenure</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Rathdown (Barony)</marc:subfield> </marc:datafield> <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land use surveys</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Wicklow (County)</marc:subfield> </marc:datafield> 11 / 13
  26. 26. Longfield Map example (Updated) <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land tenure</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Rathdown (Barony)</marc:subfield> </marc:datafield> <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land use surveys</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Wicklow (County)</marc:subfield> </marc:datafield> <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land tenure</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Rathdown (Barony)</marc:subfield> </marc:datafield> <marc:datafield tag="650" ind1="" ind2=""> <marc:subfield code="a">Land use surveys</marc:subfield> <marc:subfield code="z">Ireland</marc:subfield> <marc:subfield code="z">Wicklow (County)</marc:subfield> </marc:datafield> <marc:datafield tag="651" ind2="7" ind1=""> <marc:subfield code="2">logainm.ie</marc:subfield> <marc:subfield code="a">Rathdown</marc:subfield> <marc:subfield code="0">http://data.logainm.ie/place/283</marc:s </marc:datafield> 11 / 13
  27. 27. Demo page: http://apps.dri.ie/locationLODer 12 / 13
  28. 28. Conclusions Creation of a new Linked Data geographical Dataset Linking to other publicly available datasets Enhancing of NLI’s MARC/XML records 13 / 13
  29. 29. Conclusions Creation of a new Linked Data geographical Dataset Linking to other publicly available datasets Enhancing of NLI’s MARC/XML records Future work Improve the Silk matching rules to obtain better matching Street level matching Enhancing the NLI’s cataloguing system (VuFind) 13 / 13
  30. 30. Conclusions Creation of a new Linked Data geographical Dataset Linking to other publicly available datasets Enhancing of NLI’s MARC/XML records Future work Improve the Silk matching rules to obtain better matching Street level matching Enhancing the NLI’s cataloguing system (VuFind) Thank you! Questions? 13 / 13

×