Jennie Choi is the General Manager of Collection Information at the Met Museum. She recently took two of Wiki Education’s courses on Wikidata and reflects on her experience with the Wikimedia community in this guest blog post.
Partnering with leading members of the Wiki community has been invaluable. We’ve learned a lot from their experience and expertise, but I wanted to expand my skill set and become more self-sufficient in editing and uploading images. My contributions started modestly at edit-a-thons hosted by the museum. During these events I learned the basics about statements, properties, and references. I filled out some items where properties were missing and manually uploaded a handful of images. Given the size of our collection (500,000 records online), and the large number of new images added since the Open Access launch (roughly 60,000). I needed to learn how to make contributions more quickly and at scale.
The Metropolitan Museum began working with the Wiki community in 2017 when we launched our Open Access program. Led by our Wikimedian-in-Residence, Richard Knipel, over 375,000 public domain images were uploaded to Wikimedia Commons that year. This resulted in a huge increase in visibility for images in our collection. Wikipedia articles on a wide range of topics including Henry VIII, Vincent Van Gogh, pineapples, Down Syndrome, and the economy of Japan have all used images from The Met’s collection. We have seen incredible growth in Wikipedia page views containing our files. Between April 2017 and April 2020 views increased by 576%.
With the launch of our public API in 2018 we expanded our collaboration with the community. During an AI hackathon with Microsoft and MIT that year we worked with Wikimedian strategist Andrew Lih who worked to create a Wikidata game that allowed users to validate depicts statements generated by an AI algorithm. We have continued working with Andrew as he has led the effort to create Wikidata items for works in our collection. Because of its structured data schema, links to other items, language independence, and reliance on Wikidata by internet search engines and voice assistants, we see great benefits in extending the reach of our collection by contributing to this important resource. Most rewarding is being able to reach new audiences and seeing our objects in new contexts. With Andrew’s guidance, over 14,000 items have been created for our objects.
My first project was to enhance our existing Wikidata items wherever possible. Many of our items were missing images, dates, depicts, and other statements important for describing artworks. Using the Quick Statements tool I was able to add over 1,000 images to our public domain records. Earlier this year I mapped all our subject keywords to their corresponding Wikidata items using Open Refine and stored the Q numbers in our cataloguing database. This made it very easy to add depicts statements to over 4,000 Wikidata items where this was lacking. My next project will be to continue creating new items for works in our collection. We are fortunate to be building on Andrew’s work. He has made tremendous progress in developing a data model for the GLAM community and has also developed a crosswalk database mapping our thousands of object names to Wikidata items. While we do have a significant number of records on Wikidata, there are large areas of the collection that have not been added yet, like British satirical prints, Civil War photographs, African textiles, and our large baseball card collection, which many people probably don’t know we have. Tools like Quick Statements make creating items much easier but we still face challenges in creating items for more complex artworks. Many objects have multiple creators like the tapestry room Croome Court, which has several designers and makers who worked on portions of the room, as well as a manufacturer, and a workshop director. Most of the qualifiers for these names like “Room after a design by” and “Plaster ceiling and cornice moldings by” do not exist on Wikidata. Other objects like suits of armor may consist of different components created during different and approximate time periods. Sword scabbards can have multiple dimensions and will often include weights. More work is needed in developing a data model that will allow us to more accurately describe our collection, which we can then share with other art museums.
In addition to enhancing and creating new Wikidata items I’ve started uploading more of our public domain images to Wikimedia Commons. During the past three years we have acquired more works and digitized thousands of new objects. We want to continue sharing our open access content with the world. Using the Patty Pan tool I uploaded over 5,000 images during the past month. I’m hoping the development of new tools will make it easier to add structured data statements to our Commons images, which will make our collection even more discoverable.
We’re still in the early stages of our work with Wikidata. There are many more areas I look forward to exploring, including creating records for all our artists, using queries to generate compelling data visualizations, adding translated content from our printed guidebook, and working with other museums to further the development of data models for complex artworks. I’d like to develop a strategy to improve accessibility to our images on Commons. With hundreds of thousands of available images, how can we help users find our files more easily? I’d also like to explore ways to keep our records up to date in an automated manner. Our staff make cataloging changes everyday, how can we ensure these changes are reflected on Wikidata?
Contributing our content to Wikimedia Commons and Wikidata has allowed the Met to further fulfill our mission to connect people to creativity, knowledge, and ideas. We are seeing our works used in new contexts that go well beyond art history while hopefully creating some inspiration along the way.
Interested in taking a course like the one Jennie took? Visit learn.wikiedu.org to see current course offerings.