Wikidata:WikiProject LD4 Wikidata Affinity Group/Affinity Group Calls/Meeting Notes/2020-05-19

From Wikidata
Jump to navigation Jump to search

Call Agenda[edit]

  • Date: 2020-05-19
  • Topic: Wikidata + Education projects in Brazil--"If Wikidata is the solution...what is the need?"
  • Speaker: João Alexandre Peschanski

Presentation Materials[edit]

Meeting Notes[edit]

  • What are we trying to develop with Wikidata?
  • Bring projects led as professor, Wikimedian and researcher and try to distill understanding of how we are crossing a frontier in education--computational communications
    • Less theoretical than it is practical
    • Have strong theoretical sense of what we are doing
  • Show how we’ve been exploring Wikidata in Brazil
    • Ederporto, GiFontenelle and EricaAzzellini and others
  • Why talking about a need--not so clear what can be achieved with Wikidata?
    • Know how to achieve metrics in Wikipedia
    • With Wikidata exploring new phase of Wikimedia projects and education
  • Explore Wikidata as a solution through 3 projects
  • Wikidata has become solution to solve 3 needs
    • How can we improve efficiently and effectively content on municipal elections in Brazil on Wikipedia?
      • Good sources of information are hard for people to access and understand
    • How can we reconcile data from heterogenous databases on killed and disappeared in Brazilian military dictatorship?
      • Disagreeing data--4 databases that disagree--how to reconcile data?
    • How can we map if monuments in São Paulo are where they are claimed to be?
      • How to think of Wikidata as creating new information architecture
  • Accessory use
    • Wikidata as a resource for Wikimedia projects
    • Human editors can get bored
      • Start to make mistakes that could be prevented with use of automated resources
      • Interesting connection between bots and humans--need human editors to make reasonable decision about content
  • Structuring Wikidata with intentionality
    • Students working on articles in Wikipedia
    • Creating a template to create output of electoral results in table form in Wikipedia sandbox using data from Wikidata
      • Discussion about making process easier for users
    • Gives student structure for working on article
      • Introduction written with templates, can see infobox, tables (brought in automatically)
      • Students can work on analysis part of article now that the rest has been automated
  • Wikibook in Portuguese
    • Book created on famous photographer
    • Create digital book with images uploading
    • Creating Reasonator illustration of the page with photograph
      • Descriptive elements
      • Queries
      • Set of tools you can connect with
      • Code behind this one line--only important element is the Wikidata item--everything being transcluded from Wikidata
        • Can converge several tools that have been created by people in several countries to one place
          • Wikidata Art Explorer--can describe image with depicts statements
  • Power of Wikidata as an accessory for very cool projects
  • Question: In the election page process, is the introductory paragraph also generated from Wikidata?
    • Yes--use a structured narrative template for it
  • Do the students edit Wikidata?
    • Yes
  • How do you check any mistaken description added to an image?
    • Trust community
    • Museums check what’s added and whether
    • AI can help
  • Towards a structured Wikimedia
    • Have people work on what requires intelligence and creativity
      • Azzellini et al. (2019)
      • Vrandecic (2020)
  • Metadata curation
    • Wikidata as a solution for disagreeing data
      • Statements could all be true, so not necessarily choose one over other, but interconnect them
      • Need a language across the databases--language of reconciliation
      • Human editors reasonable--able to solve data conundra
        • This is where students are editing
    • Military dictatorship databases
      • Can completely disagree
      • Who is right the worst question to ask
      • Need to assume everyone right and create dialog
      • List generated by ListeriaBot
        • Transclusion of references--depends on database
      • Edits done by students
        • Disagreeing data--checked where information coming from
          • Is it a typo?--look at birth certificates, primary sources
          • Actual disagreement that can’t be solved--keep both data
        • Students have worked on editing the people on Wikipedia who have been killed during the dictatorship
      • Using Wikidata to create common language
    • Developed a game
      • Check the museum collection
      • How many of these things are seen in image?
        • Museum can collect information and see if what has been done as community is useful and then can bring back to local database
      • Question: Could the form for game be designed to work with Cradle and Schema (ShEx)?
        • Can follow up with Eder
    • Roundtripping--use structured data on commons with hope of eventually rountripping
      • Ex. this item in Wikicommons M74329853 same as File:Manifestação estudantil contra a Ditadura Militar 577.tif we need to define an easy way like ORCID or DOI to add the location of a picture(s) so we can easy check the M74329853 were we find other pictures and can then extracmetadata from more places t
  • Towards a meta-database
    • Data artifacts
      • Wikidata as computational media process
      • Creating what we can’t find on the internet
    • Question: What distinguishes “new digital objects” in this case from “original research”? Or are they OR? WD, like WP, discourages that: https://www.wikidata.org/wiki/Wikidata:What_Wikidata_is_not
      • Less about creating new objects, but checking investigative elements
    • Data can be visually investigated and can lead to creating new  digital objects
    • Journalism objects created based on Wikidata
      • Queries
        • City government stating where monuments located--public database
          • Could check to see if monuments were there--not stolen, moved, mistaken location
          • Student went to check that monument location was where it should be and edited in Wikidata
          • Can then use query to generate map
            • Can see irregularity of monuments across the city--cultural deserts in poorer communities
              • Not an initial thought from project, but emerged and was used in São Paulo newspapers
            • Could check how many women depicted on monument?
              • 9 of 1200 monuments depicted them
          • By using Wikidata your content can continually improve
        • Map done from Wikimedia to check dams that were at risk
          • Yellow dots considered dams that could lead to disaster
          • Green and blue okay or almost okay
          • Check each dam with students
            • Call dams and communicate with technicians
            • Some missing information--doesn’t exist
            • Investigative use of Wikidata
              • Items already exist on Wikidata
                • Done as hackathon
        • Using Wikidata as element to organize references
          • Articles used on Wikipedia in Portuguese and brought to Wikidata
          • Relies on Scholia
          • Can check most frequent keywords that journal has
          • Check images of authors and which authors published the mosts and visualize their connections
          • Had Wikidata lab on this work
        • Accessibility
          • Spoken Wikipedia
            • Through Wikidata having students create audible versions of paintings that are then associated with the Wikidata entry and eventually show up in Wikipedia infoboxes
            • Can provide very quickly a set of products that museums, NGOs could rely on
            • Language determined at this point--only have Brazilian Portuguese
            • Structure depicts so audible versions can be available in any language
    • Query based media production
      • Can distill more theoretical points
        • Structuring Wikipedia--structured gambiarra--workaround--not sure just try it
        • Interconnecting databases
        • Querying
      • Innovations in process, objects/product
      • Thinking about Wikidata as part of ecology in digital world
      • Data literacy included in academic curriculum
      • Received 300,000 images that know nothing about because were censored during military dictatorship
        • Structured data on Commons as first step to see what is in picture
        • Can eventually find when image was taken
        • Hopes will be able to query in future
    • Invitation to Wikidata Lab XXIII--Shani Evanstein--thinking about Wikidata and education connection--working on Ph.D.
    • Question: How to transclude Wikibooks from wikidata items?
    • Added to Wikibooks in Portuguese a module that allows us to create structured narrative template: https://pt.wikibooks.org/wiki/M%C3%B3dulo:Mbabel
    • Question: What advice do you have for people hosting wikidata hackathons for undergraduate students?
      • Think about need--less about learning technology--don’t make good use of technology they know how to use
      • Need that they can solve--empowerment
      • Could be more meaningful as framing as solving issue--festival--people engage with project because of making a difference in world--not based around technology
      • Make sure that doesn’t get boring--need to automate part of the work, so can build upon that
      • Development hackathon--empowering student to present development
        • Wikidata Labs usually very technical--choose an undergrad or recent grad student to present--have 3-4 months to prepare presentation--everyone can learn it and learn to the point where you can collaboratively share it
        • Tec
    • Question: is it possible to automate data insertion using some technology together with curriculo lattes?
      • Academic curriculum database in Brazil--should be able to provide in CSV, but doesn’t work
      • Should be possible--data is available, but there are some political barriers
      • Requires real technical investment--don’t make easy for automatic scraping
      • Often scrape databases--contact institutions if cannot get through the government