Upgrade to Pro — share decks privately, control downloads, hide ads and more …

spatial@linkedscience – Exploring the Research Field of GIScience with Linked Data

Ee36c21b1a92a643c73b120fafe10b54?s=47 Carsten Keßler
September 19, 2012

spatial@linkedscience – Exploring the Research Field of GIScience with Linked Data

Presentation of our paper at GIScience 2012 in Columbus, OH. PDF of the paper: http://carsten.io/GIScience2012.pdf

Ee36c21b1a92a643c73b120fafe10b54?s=128

Carsten Keßler

September 19, 2012
Tweet

Transcript

  1. Carsten Keßler 1, Krzysztof Janowicz 2, Tomi Kauppinen 1 1

    Institute for Geoinformatics | University of Münster, Germany 2 Department of Geography | University of California | Santa Barbara, USA http://carsten.io | @carstenkessler <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12/authors0>  rdf:rest  <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12/authors01>  . <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12/authors01>  a  rdf:List  . <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12/authors01>  rdf:first  <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  . <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12/authors01>  rdf:rest  rdf:nil  . <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  foaf:publications  <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12>  . <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  foaf:publications  <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12>  . <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  foaf:publications  <http://spatial.linkedscience.org/context/giscience/paper/doi10.1007/3-­‐540-­‐45799-­‐2_12>  . <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  a  foaf:Person  . <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  foaf:name  "Maozhen  Li"  . <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  foaf:givenName  "Maozhen"  . <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  foaf:familyName  "Li"  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  foaf:member  <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  . <http://spatial.linkedscience.org/context/giscience/membership11>  a  rdfs:Statement  . <http://spatial.linkedscience.org/context/giscience/membership11>  rdfs:subject  <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  . <http://spatial.linkedscience.org/context/giscience/membership11>  rdfs:predicate  foaf:member  . <http://spatial.linkedscience.org/context/giscience/membership11>  rdfs:object  <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  . <http://spatial.linkedscience.org/context/giscience/membership11>  dc:date  "2002"  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  a  foaf:Organization  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  foaf:name  "Cardiff  University  Deptartment  of  Computer  Science  Cardiff  CF24  3XF  UK"  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  <http://lodum.de/helper/addressType>  "locality"  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  <http://lodum.de/helper/addressType>  "political"  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  vcard:ADR  "Cardiff,  UK"  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  geo:lat  51.4815810  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  geo:long  -­‐3.1790900  . <http://spatial.linkedscience.org/context/affiliation/affiliationc36035e4c2fcc19f0bab56ad5838ec24>  geo:lat_long  "51.4815810  -­‐3.1790900"  . <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  foaf:knows  <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  . <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  foaf:knows  <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  . <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  foaf:knows  <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  . <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  foaf:knows  <http://spatial.linkedscience.org/context/person/person690da4aa2199a272385499a8c7305eb9>  . <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  a  foaf:Person  . <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  foaf:name  "Sheng  Zhou"  . <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  foaf:givenName  "Sheng"  . <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  foaf:familyName  "Zhou"  . <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  foaf:knows  <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  . <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  foaf:knows  <http://spatial.linkedscience.org/context/person/persondcb5f05cab6c31d6d00bb13005b95ea3>  . <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  a  foaf:Person  . <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  foaf:name  "Christopher  B.  Jones"  . <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  foaf:givenName  "Christopher  B."  . <http://spatial.linkedscience.org/context/person/persone6c3f8abba6c51d2cfcc09600b3470f9>  foaf:familyNa Spatial@LinkedScience Exploring the Research Field of GIScience with Linked Data
  2. What is Linked Science? ‣ Connecting and semantically annotating scientific

    resources: ‣ workflows, processes, models, data, methods, tools, software environments, evaluation metrics, … ‣ Needs: ‣ shared conceptualizations, ‣ well-defined ontologies and vocabularies, and ‣ reasoning mechanisms to retrieve, evaluate, and transfer scientific knowledge. Tomi Kauppinen, Alkyoni Baglatzi and Carsten Keßler (forthcoming 2012) Linked Science: Interconnecting Scientific Assets. In Terence Critchlow and Kerstin Kleese-Van Dam (Eds.): Data Intensive Science. CRC Press, USA.
  3. Linked (Open) Science Principles ‣ Scientific data is published on

    the Web as Linked Data ‣ Implementations of methods are published as open source, and methods make use of Linked Data ‣ Cloud computing is used to support running of methods by (basically) anyone ‣ Licenses and copyrights about scientific resources in use are made explicit
  4. Publications as the “Hook” ‣ Papers act as entry points

    ‣ Make the metadata linkable ‣ URLs as unique and resolvable IDs for papers, people, institutions, …
  5. Publications as the “Hook” ‣ Papers act as entry points

    Paper ‣ Make the metadata linkable ‣ URLs as unique and resolvable IDs for papers, people, institutions, …
  6. Publications as the “Hook” ‣ Papers act as entry points

    Paper ‣ Make the metadata linkable ‣ URLs as unique and resolvable IDs for papers, people, institutions, … Data Models Software Projects …
  7. Publications as the “Hook” ‣ Papers act as entry points

    Paper ‣ Make the metadata linkable ‣ URLs as unique and resolvable IDs for papers, people, institutions, … Data Models Software Projects … ‣ not possible with CiteSeer, Springerlink, DBLP, Google Scholar, ACM library…
  8. spatial.linkedscience.org ‣ Provides these hooks for the GIScience community ‣

    Source data: BibTex files from Springer and ACM
  9. @incollection  {springerlink:10.1007/978-­‐3-­‐642-­‐33024-­‐7_8,  author  =  {Keßler,  Carsten  and  Janowicz,  Krzysztof  and

     Kauppinen,  Tomi},  year  =  {2012},      title  =  {spatial@linkedscience  –  Exploring  the  Research  Field  of   GIScience  with  Linked  Data},      affiliation  =  {Institute  for  Geoinformatics,  University  of  Münster,   Germany},            booktitle  =  {Geographic  Information  Science},      series  =  {Lecture  Notes  in  Computer  Science},      editor  =  {Xiao,  Ningchuan  and  Kwan,  Mei-­‐Po  and  Goodchild,  Michael  and   Shekhar,  Shashi},      publisher  =  {Springer  Berlin  /  Heidelberg},      isbn  =  {978-­‐3-­‐642-­‐33023-­‐0},      keyword  =  {Computer  Science},      pages  =  {102-­‐115},      volume  =  {7478},      url  =  {http://dx.doi.org/10.1007/978-­‐3-­‐642-­‐33024-­‐7_8},      note  =  {10.1007/978-­‐3-­‐642-­‐33024-­‐7_8},      abstract  =  {Metadata  for  scientific  publications  contain  various   explicit  and  implicit  spatio-­‐temporal  references.  Data  on  conference   locations  as  well  as  author  and  editor  affiliations  –  both  changing  over   time  –  enable  insights  into  the  geographic  distribution  of  scientific   fields  and  particular  specializations.  At  the  same  time,  these  byproducts   of  scientific  bibliographies  offer  a  great  opportunity  to  integrate  data   across  different  bibliographies  to  get  a  more  complete  picture  of  a   domain.  In  this  paper,  we  demonstrate  how  the  Linked  Data  paradigm  can  
  10. Data Conversion

  11. Data Conversion ‣ Generating unique identifiers

  12. Data Conversion ‣ Generating unique identifiers ‣ Paper pattern: http://spatial.linkedscience.org/context/conference/

    paper/doiDOI
  13. Data Conversion ‣ Generating unique identifiers ‣ Paper pattern: http://spatial.linkedscience.org/context/conference/

    paper/doiDOI ‣ Example: http://spatial.linkedscience.org/context/acmgis/ paper/doi10.1145/1653771.1653787
  14. Data Conversion ‣ Generating unique identifiers ‣ Paper pattern: http://spatial.linkedscience.org/context/conference/

    paper/doiDOI ‣ Example: http://spatial.linkedscience.org/context/acmgis/ paper/doi10.1145/1653771.1653787 ‣ Vocabularies: Dublin Core, Bibo, Friend Of A Friend, VCard, W3C Geo
  15. Geocoding affiliations

  16. Geocoding affiliations ‣ Google Geocoding API

  17. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found
  18. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards
  19. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards
  20. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards
  21. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards
  22. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards
  23. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards
  24. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards
  25. Geocoding affiliations ‣ Google Geocoding API ‣ Iterative shortening of

    strings until result is found ‣ Example: Fantastic Research Lab, Department of Geostuff, University of Hogwards ✔
  26. Reconciliation

  27. Reconciliation ‣ Persons: James A. Hendler vs Jim Hendler

  28. Reconciliation ‣ Persons: James A. Hendler vs Jim Hendler ‣

    Affiliations: University of Hogwards vs Hogwards University
  29. Reconciliation ‣ Persons: James A. Hendler vs Jim Hendler ‣

    Affiliations: University of Hogwards vs Hogwards University ‣ Generation of same-as links based on:
  30. Reconciliation ‣ Persons: James A. Hendler vs Jim Hendler ‣

    Affiliations: University of Hogwards vs Hogwards University ‣ Generation of same-as links based on: ‣ String similarity
  31. Reconciliation ‣ Persons: James A. Hendler vs Jim Hendler ‣

    Affiliations: University of Hogwards vs Hogwards University ‣ Generation of same-as links based on: ‣ String similarity ‣ Spatial distance
  32. Reconciliation ‣ Persons: James A. Hendler vs Jim Hendler ‣

    Affiliations: University of Hogwards vs Hogwards University ‣ Generation of same-as links based on: ‣ String similarity ‣ Spatial distance ‣ SILK Framework
  33. Stats ‣ 1305 papers ‣ 2346 authors ‣ 71002 triples

  34. Stats ‣ 1305 papers ‣ 2346 authors ‣ 71002 triples

    COSIT 331 AGILE 139 ACM GIS 699 GIScience 136 Papers/Conference
  35. Stats ‣ 1305 papers ‣ 2346 authors ‣ 71002 triples

    0 100 200 300 400 1992 1993 1995 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 COSIT 331 AGILE 139 ACM GIS 699 GIScience 136 Papers/Conference
  36. None
  37. Screenshot: Paper

  38. None
  39. Complex queries ‣ http://spatial.linkedscience.org/sparql ‣ Which authors have co-authored more

    than 2 papers together? ‣ Which topics were popular at which times? ‣ Is there a trend in that certain universities preferably hire from certain other universities? ‣ …
  40. Example: Bridge-builders Benjamin Adams Christophe Claramunt (a) Matt Duckham Max

    J. Egenhofer Leila De Floriani Andrew U. Frank (a) Mark Gahegan Krzysztof Janowicz (a) Christopher B. Jones Lars Kulik Kai-Florian Richter Claus Rinner Andrea Rodríguez John Stell Egemen Tanin Stephan Winter (a) Michael Worboys
  41. The Platform ‣ Twitter Bootstrap & jQuery talking to SPARQL

    endpoint <!DOCTYPE html> <html> <!-- created 2010-01-01 --> <head> <title>sample</title> </head> <body> <p>Voluptatem 

accusantium totam 

rem 

aperiam.</p> </body> </html> HTML HTML HTTP Server Static website layout SPARQL query results via AJAX SPARQL Endpoint Triple Store Server Client
  42. Limitations ‣ Incomplete input data -> affiliations ‣ Limits de-duplication

    based on space/time ‣ Keywords are fairly useless so far ‣ Geocoding is limited by the API ‣ Dataset gives an incomplete picture of the field so far ‣ AJAX approach does not scale well
  43. What’s next?

  44. What’s next? ‣ More conferences, add journals

  45. What’s next? ‣ More conferences, add journals ‣ Improve geocoding

  46. What’s next? ‣ More conferences, add journals ‣ Improve geocoding

    ‣ Complete outlinking to GeoNames
  47. What’s next? ‣ More conferences, add journals ‣ Improve geocoding

    ‣ Complete outlinking to GeoNames ‣ Support for spatial queries via GeoSPARQL
  48. What’s next? ‣ More conferences, add journals ‣ Improve geocoding

    ‣ Complete outlinking to GeoNames ‣ Support for spatial queries via GeoSPARQL ‣ Cross-reference other sources such as LinkedDBLP
  49. Conclusions

  50. Conclusions ‣ Long-term goal: one stop shop for publications from

    the GIScience community
  51. Conclusions ‣ Long-term goal: one stop shop for publications from

    the GIScience community ‣ Lots of interesting questions to ask
  52. Conclusions ‣ Long-term goal: one stop shop for publications from

    the GIScience community ‣ Lots of interesting questions to ask ‣ Great potential in spatio-temporal properties to support reconciliation
  53. Conclusions ‣ Long-term goal: one stop shop for publications from

    the GIScience community ‣ Lots of interesting questions to ask ‣ Great potential in spatio-temporal properties to support reconciliation ‣ spatial.linkedscience.org is just one application using this dataset
  54. Conclusions ‣ Long-term goal: one stop shop for publications from

    the GIScience community ‣ Lots of interesting questions to ask ‣ Great potential in spatio-temporal properties to support reconciliation ‣ spatial.linkedscience.org is just one application using this dataset ‣ github.com/crstn/spatial-­‐linkedscience
  55. Thank you! carsten.kessler@uni-muenster.de | http://carsten.io | @carstenkessler Carsten Keßler |

    Krzystof Janowicz | Tomi Kauppinen We’re hiring!