Linked Data Survey results 1 – Who’s doing it (Updated)

LOD_Cloud_Diagram_as_of_September_2011 wikimedia.orgOCLC Research conducted an international linked data survey for implementers between 7 July and 15 August 2014. This is the first post in the series reporting the results. 

We received 122 responses to the international linked data survey conducted between 7 July and 15 August 2014. (So who is using linked data? And for what?) OCLC colleagues also responded to the survey and reported on 6 projects/services. So now we have some answers to share!

Although the survey was designed specifically for implementers of linked data projects/services, 26 of the 122 responses said they had not implemented nor were implementing a linked data project. Seven of them plan on implementing a linked data project within the next two years and 10 are planning to apply for funding to implement one. Some of them also pointed to interesting linked data projects they’re tracking, which included respondents to the survey (Oslo Public Library, BIBFRAME, Europeana, Yale Center for British Art).

The remaining 96 responses reported implementing 172 linked data projects/services; 76 of them are described. 25 of the projects consume linked data; 4 publish linked data; 47 both consume and publish linked data.  We have a good international representation. Just over half are linked data projects/services in the US, but the rest are from 14 countries: Australia, Canada, Czech Republic, France, Germany, Ireland, Italy, The Netherlands, Norway, Singapore, South Korea, Spain, Switzerland, and the UK. I’ve appended the list of the 47 institutions that took the time to describe their linked data projects/services below (several described multiple projects).

Of the 76 projects:

27 are not yet in production;
13 have been in production for less than one year;
12 have been in production for more than one year but less than two years;
24 have been in production for more than two years.

Four of the projects are “private”, for that institution’s use only. Most projects/services that have been implemented have been receiving an average of fewer than 1,000 requests a day over the last six months. The most heavily used linked data datasets as measured by average number of requests a day:

  • OCLC’s WorldCat.org  with an average of 16 million requests/day (all WorldCat.org pages include linked data)
  • Library of Congress’ id.loc.gov with over 100,000 requests/day (also the linked data source used by most of the respondents)
  • OCLC’s VIAF, also with over 100,000 requests/day
  • OCLC’s WorldCat.org Works, also with over 100,000 requests/day
  • American Numismatic Society’s thesaurus of numismatic concepts used by archeological projects and museum databases with 10,000 – 50,000 requests/day
  • British Library’s British National Bibliography, also with 10,000 – 50,000 requests/day
  • OCLC’s Dewey, also with 10,000 – 50,000 requests/day (but reflects only HTML pages served, and not requests to the RDF data).

Since so many projects have not been implemented or implemented relatively recently, only 47 could assess whether the linked data project/service was successful in achieving its desired outcome. Most (33) said it was successful or mostly successful.  Success measures included “increasing international use”, improved workflow, moving from a pilot to production, ability to link data by relationships, just making the data available as linked data, and professional development of staff. Several noted the need for better metrics to assess the service value, the challenges of harmonizing the data, and the lack of identifiers to link headings to.

Parts of institution involved: Most of the respondents reported that multiple units of the institution were involved in their linked data project/service. When only one part of an institution was involved, it was more likely the research and development group (8 projects, but cited in 19 projects total.)  Library and/or archives were involved the most, cited as being involved in 52 projects. Metadata services was the next most-involved, in 37 projects. Digital library services and library systems/IT or campus IT were involved in a third or more of the projects. Seventeen involved digital humanities and/or faculty in academic departments. The University College Dublin’s Digital Library involved the most units of all the described projects, with 10: library, archives, metadata services, digital library services, library systems/information technology, research and development group, computer science department, digital humanities, campus museum, faculty in academic departments.

External groups involved: Seventeen of the projects did not involve any external groups or organizations.  Fifteen were part of a national and/or international collaboration. Twenty-five involved other libraries or archives and 19 other universities or research institutions. Thirteen involved a systems vendor or a corporation/company.  Nine collaborated with other consortium members; eight were part of a discipline-specific collaboration. Europeana listed the most external groups, with 6: other libraries/archives, other universities/research institutions, other members of their consortium, part of a discipline-specific collaboration, part of an international collaboration, and a large network of experts working in cultural heritage.

Staffing:  Almost all of the institutions that have implemented or are implementing linked data projects/services have added linked data to the responsibilities of current staff (61); only 12 have not. Thirteen have staff dedicated to linked data projects (five of them in conjunction with adding linked data to the responsibilities of current staff). Five are adding or have added new staff with linked data expertise; ten are adding or have added temporary staff with linked data expertise; and seven are hiring or have hired external consultants with linked data expertise.

Funding: Twenty-nine of the projects received grant funding to implement linked data; most projects (51) were covered by the library/archive or the parent institution. Three linked data projects received funding support from partner institutions; four linked data projects were privately funded.

Linked data survey respondents describing linked data projects/services

  • American Antiquarian Society
  • American Numismatic Society
  • Archaeology Data Service (UK)
  • Biblioteca della Camera dei deputati (Italy)
  • British Library
  • British Museum
  • Carleton College
  • Charles University in Prague
  • Colorado College
  • Colorado State University
  • Cornell University
  • Data Archiving and Networked Services, Royal Netherlands Academy of Arts and Sciences
  • Digital Public Library of America
  • Europeana Foundation
  • Fundacción Ignacio Larramendi (Spain)
  • Goldsmiths’ College
  • Library of Congress
  • Minnesota Historical Society
  • Missoula Public Library
  • National Library Board (NLB) of Singapore
  • National Library of Medicine
  • North Carolina State University Libraries
  • NTNU (Norwegian University of Science and Technology) University Library
  • OCLC
  • Oslo Public Library
  • Public Record Office, Victoria
  • Queen’s University Library
  • Research Libraries UK
  • Smithsonian
  • Springer
  • Stanford University
  • Stichting Bibliotheek.nl
  • The European Library
  • The University of Texas at Austin
  • Tresoar (Leeuwarden – The Netherlands)
  • University College Dublin
  • University College London (UCL)
  • University of Alberta Libraries
  • University of Bergen Library
  • University of British Columbia
  • University of California-Irvine
  • University of Illinois at Urbana-Champaign
  • University of North Texas
  • University of Oxford
  • University of Pennsylvania Libraries
  • Western Michigan University
  • Yale Center for British Art

[Originally posted 2014-08-28, updated 2014-09-04]

5 Comments on “Linked Data Survey results 1 – Who’s doing it (Updated)”

  1. The 4 September update includes Research Libraries UK and The European Library. The statistics were all updated, and also now include the responses from OCLC’s six projects.

  2. There are others of significance that aren’t listed here, but may be elsewhere in the report (?).

    * The National Library of Sweden’s LIBRIS system. It was the topic of an LC webcast:
    http://www.loc.gov/today/cyberlc/feature_wdesc.php?rec=5521

    * The FAO Agris catalog and AGROVOC vocabulary were among the first to use linked data. http://aims.fao.org/openagris

    * The National Library of Spain provides linked data: http://www.semantic-web-journal.net/content/datosbnees-library-linked-data-dataset

    * The DataHub has 18 library linked data datasets http://datahub.io/group/lld

Comments are closed.