Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo

Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums

•
•
PACKED vzw
PACKED vzw

In order for museums to truly reap the benefits of publishing their collections online in a sustainable way, PACKED vzw presents the results of its Linked open data project as a best practice guide for the Flemish heritage sector.

1 of 50
Linked Open Data publications through Wikidata & persistent identification in Flemish museums
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Museum as an online knowledge source ‘Offering original authentic information from trusted sources’: • permanently available • access to authentic and up-to-date information about artworks • easy to find and to reuse
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Persistent URI’s
artwork http://cvg. be/collectie/work/i d/1856 data http://kmska. be/collectie/work/ data/1856 picture http://kmska. be/collectie/work/ representation/185 6 picture http://lukasweb. be/collectie/work/ representation/185 6 data http://cvg. be/collectie/work/ data/1856 creator http://viaf.org/ viaf/312406452/ type http://vocab.getty. edu/aat/3000337 99
DATA CLUSTERS
RESOLVER Open source! https://github.com/PACKED-vzw/resolver
RESOLVER API / collectie- beheersysteem DAM - systeem resolver CMS collectie website Data manager europeanamobiele apps museum website URI URL (JPEG) URL(XML) URL (HTML) URI URI
https://www.projectcest.be/images/3/38/20150430_Handout_Open_Refine_workshop.pdf
ENRICHMENT THROUGH LINKED DATA External authorities: - Wikidata, VIAF, RKDartists, ODIS, Geonames, AAT (Getty Vocabularies), Iconclass https://www.projectcest.be/wiki/Publicatie: Project_Persistente_identificatie
More contextual data!
Event-based object description Wikidata, VIAF, RKDartists.. Event 1 Event 2 Event 3 Event 4 Event 5 Geonames, TGN … Wikidata; Worldcat; Amazon api; Librarything api; LoC CDS; Olid; Openlibrary.org… ISO 8601
Lifecycle of an artwork What happend to an artwork over time and where did this event take place?
Lifecycle of an artwork What happend to an artwork over time and where did this event take place?
Lifecycle of an artwork What happend to an artwork over time and where did this event take place?
Provenance timeline How did an artwork change owners through time?
Exhibitions map In which countries were artworks from a particular collection shown in an exhibition?
Acquisition source From which sources did a museum acquire artworks? (one museum)
Acquisition source From which source did a museum acquire artworks? (five museums)
Objectname/Iconography streamgraph How important was an objectname in a collection through time? https://www.projectcest.be/wiki/Publicatie:Event-based_objectbeschrijvingen
Museums & Linked Open Data via Wikidata
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
What were we going to do? 1. PACKED writes a whitepaper on Wikidata (how can you add data to Wikidata, what are Wikidata advantages to your institution, # best practices). 2. Project partners send a CSV dataset to us containing a.o. the created PIDs for their artworks. 3. This dataset is uploaded to Wikidata under CC0 license - the info from the set is then used in different Wiki-channels (a.o. Wikipedia) 4. Partners get back an RDF export of the data they delivered to us
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Why would you, as a museum, participate? 1. Because it is done for FREE 2. PACKED & Wiki-community do the work: MINIMAL INPUT 3. Wikidata feeds Wikipedia = REACH (BIGGER/OTHER/...) AUDIENCE 4. Lots of possibilities for REUSE in your own projects, cf. by means of RDF export
Our biggest help: the Wiki(data) community
Some great people here Sandra Fauconnier, hired as project staff member Foto: Eneasmx, CC-BY-SA 4.0 Maarten Dammers, volunteer data uploader Foto: Sarah Stierch, CC-BY 4.0
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
We made a nice explanatory screencast Publicly available via YouTube / only in Dutch
We wrote a whitepaper Chapters 1. Wikimedia, Wikipedia, Wikidata 2. Costs and benefits of using Wikidata (=business cases) 3. Crosswalk, delivery, upload process 4. Conclusion Annex: cases usage guidelines
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Business cases / costs and benefits Museums Wikimedia Society Benefits 1. Low cost 2. Audience reach 3. Creativity 4. Data in wider context 1. Mission accomplished 2. Quality data 3. Time to learn 4. Incremental growth of data 1. Heritage visible to audience 2. Low cost 3. Open data 4. Source for Wikipedia articles Costs 1. Waive exclusivity 2. Time invested for maintenance 3. Time invested for cleaning 1. Cost for storage 2. Time investment (volunteers) 3. Need for new tools
Museums sent us their dataset ... What was included? ● Results from previous PID project ● Almost no ‘own’ information ● Largely external enrichments ● Inluding persistent resolverlinks
… this got uploaded ... Largest load: through bot by volunteer Maarten (Multichill)
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
… and the CC0 license got ok-ed The project page of Wiki Loves Open Data explains the expectations of the Wikidata community regarding donated open data:
Information = facts. You ONLY provide the information from your CSV set, NOT the corresponding digital images. What WILL be on Wikidata: ● Persistent links going to your resolver ● Links referring to other external authorities, like e.g. VIAF The ‘own’ data you provide under CC0 is title of the artwork, creator, date(s), object name/type and the name of your organisation. As CC0 permits any kind of reuse, you can add Usage Guidelines. Though not legally binding, they may touch on the ‘good family father’ principle. MoMa, Tate and Cooper- Hewitt have done this too.
NOTABLE (‘relevance’) Wikidata: 2 goals ● Wikidata centralises links between different language versions in Wikimedia projects ● Wikidata serves as a generic knowledge bank for the entire world In there: 3 criteria (at least 1 should be met) 1. The item contains at least one valid sitelink to a content page on a Wikimedia project. 2. The item refers to a clearly identifiable concept or material entity, and is sufficiently relevant, which can be proven by the fact that the item is described in serious, public sources. 3. The item covers a structural need; for example it is useful for describing other items. The Reading by Emile Verhaeren (Théo van Rysselberghe) 0 Wikipedia articles but: OK! Kazuifel with saints (Anonymous, 16th century) … ?
EDITABLE, MAINTAINED ● Who are the Wikidata volunteers? ● Museums as authorities ● Wikidata -> references/sources ● There is room for a different opinion … ● … with authoritative sources! Authority control (literally) on Wikidata Peter Paul Rubens is linked to the following identifiers on Wikidata: Library of Congress Authors, VIAF, ISNI, GND, Freebase, ULAN, Bibliothèque Nationale de France, CANTIC, BBC Your Paintings, RKDartists, Sandrart.net, Oxford Biography Index, NTA, Digitale Bibliotheek der Nederlandse Letteren, EMLO, genealogics.org, Smithsonian American Art Museum, National Portrait Gallery, Web Gallery of Art, BALat, SUDOC, Artsy, Open Library, NGA, NNDB.
The numbers Wikidata now (end if Jan ‘16) contains descriptions/records on 26.680 works of art from the museums that participated in this project. At project start (Oct ‘15), there were 65. (Wikidata had 10.453 items from Dutch collections.) ● Those 26.680 works of art were created by 3.615 artists. Of these, a few hundreds are newly entered in Wikidata because of the project. ● There was a mapping of 399 different object names/types to the ‘instance of’ terms of Wikidata - from acrylicschilderij to zwart-witfoto.
The publications Whitepaper - but also a Manual on how to add Wikidata records from scratch or update/add more information to existing ones ● manually, using Quickstatements language (or not) ● using a bot (and a volunteer) Full doc contains screenshots, step-by-step design & is translated to English & will be made available on Wikidata project page (+ communication)
Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums
Closing project workshop 1. Get your own account! 2. Search for your own collection & museum on Wikimedia Commons and add an image to a Wikidata record 3. Find an item from your collection on Wikidata that had already been worked on by a volunteer a. Win a biscuit if you find a huge ‘mistake’ b. Correct / add information where needed 4. Add a brand new record to Wikidata 5. How does the watchlist work? 6. Look at the RDF version of one of the works of art from your collection. Which information do you find most interesting?
Next steps: PACKED vzw & Wikidata ● Publication of all materials in English on project page: model/inspiration for alike projects or museum communities ● Adding information to creators: meat to bones (birth/dying date/profession/authority links, …) ● Add and correct dates, match ‘parts of’ works with the sum of parts (e.g. tryptich) ● Complement materials and ‘instance of’ types ● Correct PIDlink mappings ● Check and correct doubles, wrong attributions, copies, … ● Framework of larger project: Datahub for Museums
Next steps: Museum partners ● Public side of the project: interested in Sum of all Paintings, edit-a-thons, Wiki loves Art ● Own communication! ● (Prototyping)project based on Wikidata (catalogue, app, game …) as a developing testbed or case And … an IMAGEdonation
~~ The End ~~ Questions/experiences on persistent URI, events-data or Wikidata? Share them with ALINA@PACKED.BE BARBARA@PACKED.BE

More Related Content

Linked Open Data Publications through Wikidata & Persistent Identification in Flemish Museums

  • 1. Linked Open Data publications through Wikidata & persistent identification in Flemish museums
  • 3. Museum as an online knowledge source ‘Offering original authentic information from trusted sources’: • permanently available • access to authentic and up-to-date information about artworks • easy to find and to reuse
  • 10. RESOLVER API / collectie- beheersysteem DAM - systeem resolver CMS collectie website Data manager europeanamobiele apps museum website URI URL (JPEG) URL(XML) URL (HTML) URI URI
  • 12. ENRICHMENT THROUGH LINKED DATA External authorities: - Wikidata, VIAF, RKDartists, ODIS, Geonames, AAT (Getty Vocabularies), Iconclass https://www.projectcest.be/wiki/Publicatie: Project_Persistente_identificatie
  • 14. Event-based object description Wikidata, VIAF, RKDartists.. Event 1 Event 2 Event 3 Event 4 Event 5 Geonames, TGN … Wikidata; Worldcat; Amazon api; Librarything api; LoC CDS; Olid; Openlibrary.org… ISO 8601
  • 15. Lifecycle of an artwork What happend to an artwork over time and where did this event take place?
  • 16. Lifecycle of an artwork What happend to an artwork over time and where did this event take place?
  • 17. Lifecycle of an artwork What happend to an artwork over time and where did this event take place?
  • 18. Provenance timeline How did an artwork change owners through time?
  • 19. Exhibitions map In which countries were artworks from a particular collection shown in an exhibition?
  • 20. Acquisition source From which sources did a museum acquire artworks? (one museum)
  • 21. Acquisition source From which source did a museum acquire artworks? (five museums)
  • 22. Objectname/Iconography streamgraph How important was an objectname in a collection through time? https://www.projectcest.be/wiki/Publicatie:Event-based_objectbeschrijvingen
  • 23. Museums & Linked Open Data via Wikidata
  • 25. What were we going to do? 1. PACKED writes a whitepaper on Wikidata (how can you add data to Wikidata, what are Wikidata advantages to your institution, # best practices). 2. Project partners send a CSV dataset to us containing a.o. the created PIDs for their artworks. 3. This dataset is uploaded to Wikidata under CC0 license - the info from the set is then used in different Wiki-channels (a.o. Wikipedia) 4. Partners get back an RDF export of the data they delivered to us
  • 28. Why would you, as a museum, participate? 1. Because it is done for FREE 2. PACKED & Wiki-community do the work: MINIMAL INPUT 3. Wikidata feeds Wikipedia = REACH (BIGGER/OTHER/...) AUDIENCE 4. Lots of possibilities for REUSE in your own projects, cf. by means of RDF export
  • 29. Our biggest help: the Wiki(data) community
  • 30. Some great people here Sandra Fauconnier, hired as project staff member Foto: Eneasmx, CC-BY-SA 4.0 Maarten Dammers, volunteer data uploader Foto: Sarah Stierch, CC-BY 4.0
  • 32. We made a nice explanatory screencast Publicly available via YouTube / only in Dutch
  • 33. We wrote a whitepaper Chapters 1. Wikimedia, Wikipedia, Wikidata 2. Costs and benefits of using Wikidata (=business cases) 3. Crosswalk, delivery, upload process 4. Conclusion Annex: cases usage guidelines
  • 35. Business cases / costs and benefits Museums Wikimedia Society Benefits 1. Low cost 2. Audience reach 3. Creativity 4. Data in wider context 1. Mission accomplished 2. Quality data 3. Time to learn 4. Incremental growth of data 1. Heritage visible to audience 2. Low cost 3. Open data 4. Source for Wikipedia articles Costs 1. Waive exclusivity 2. Time invested for maintenance 3. Time invested for cleaning 1. Cost for storage 2. Time investment (volunteers) 3. Need for new tools
  • 36. Museums sent us their dataset ... What was included? ● Results from previous PID project ● Almost no ‘own’ information ● Largely external enrichments ● Inluding persistent resolverlinks
  • 37. … this got uploaded ... Largest load: through bot by volunteer Maarten (Multichill)
  • 40. … and the CC0 license got ok-ed The project page of Wiki Loves Open Data explains the expectations of the Wikidata community regarding donated open data:
  • 41. Information = facts. You ONLY provide the information from your CSV set, NOT the corresponding digital images. What WILL be on Wikidata: ● Persistent links going to your resolver ● Links referring to other external authorities, like e.g. VIAF The ‘own’ data you provide under CC0 is title of the artwork, creator, date(s), object name/type and the name of your organisation. As CC0 permits any kind of reuse, you can add Usage Guidelines. Though not legally binding, they may touch on the ‘good family father’ principle. MoMa, Tate and Cooper- Hewitt have done this too.
  • 42. NOTABLE (‘relevance’) Wikidata: 2 goals ● Wikidata centralises links between different language versions in Wikimedia projects ● Wikidata serves as a generic knowledge bank for the entire world In there: 3 criteria (at least 1 should be met) 1. The item contains at least one valid sitelink to a content page on a Wikimedia project. 2. The item refers to a clearly identifiable concept or material entity, and is sufficiently relevant, which can be proven by the fact that the item is described in serious, public sources. 3. The item covers a structural need; for example it is useful for describing other items. The Reading by Emile Verhaeren (Théo van Rysselberghe) 0 Wikipedia articles but: OK! Kazuifel with saints (Anonymous, 16th century) … ?
  • 43. EDITABLE, MAINTAINED ● Who are the Wikidata volunteers? ● Museums as authorities ● Wikidata -> references/sources ● There is room for a different opinion … ● … with authoritative sources! Authority control (literally) on Wikidata Peter Paul Rubens is linked to the following identifiers on Wikidata: Library of Congress Authors, VIAF, ISNI, GND, Freebase, ULAN, Bibliothèque Nationale de France, CANTIC, BBC Your Paintings, RKDartists, Sandrart.net, Oxford Biography Index, NTA, Digitale Bibliotheek der Nederlandse Letteren, EMLO, genealogics.org, Smithsonian American Art Museum, National Portrait Gallery, Web Gallery of Art, BALat, SUDOC, Artsy, Open Library, NGA, NNDB.
  • 44. The numbers Wikidata now (end if Jan ‘16) contains descriptions/records on 26.680 works of art from the museums that participated in this project. At project start (Oct ‘15), there were 65. (Wikidata had 10.453 items from Dutch collections.) ● Those 26.680 works of art were created by 3.615 artists. Of these, a few hundreds are newly entered in Wikidata because of the project. ● There was a mapping of 399 different object names/types to the ‘instance of’ terms of Wikidata - from acrylicschilderij to zwart-witfoto.
  • 45. The publications Whitepaper - but also a Manual on how to add Wikidata records from scratch or update/add more information to existing ones ● manually, using Quickstatements language (or not) ● using a bot (and a volunteer) Full doc contains screenshots, step-by-step design & is translated to English & will be made available on Wikidata project page (+ communication)
  • 47. Closing project workshop 1. Get your own account! 2. Search for your own collection & museum on Wikimedia Commons and add an image to a Wikidata record 3. Find an item from your collection on Wikidata that had already been worked on by a volunteer a. Win a biscuit if you find a huge ‘mistake’ b. Correct / add information where needed 4. Add a brand new record to Wikidata 5. How does the watchlist work? 6. Look at the RDF version of one of the works of art from your collection. Which information do you find most interesting?
  • 48. Next steps: PACKED vzw & Wikidata ● Publication of all materials in English on project page: model/inspiration for alike projects or museum communities ● Adding information to creators: meat to bones (birth/dying date/profession/authority links, …) ● Add and correct dates, match ‘parts of’ works with the sum of parts (e.g. tryptich) ● Complement materials and ‘instance of’ types ● Correct PIDlink mappings ● Check and correct doubles, wrong attributions, copies, … ● Framework of larger project: Datahub for Museums
  • 49. Next steps: Museum partners ● Public side of the project: interested in Sum of all Paintings, edit-a-thons, Wiki loves Art ● Own communication! ● (Prototyping)project based on Wikidata (catalogue, app, game …) as a developing testbed or case And … an IMAGEdonation
  • 50. ~~ The End ~~ Questions/experiences on persistent URI, events-data or Wikidata? Share them with ALINA@PACKED.BE BARBARA@PACKED.BE