Giovanna Fontenelle (she/her) Program Officer; GLAM and Culture |
Several product teams at the Foundation are working hard to improve image discovery and reuse on Wikimedia projects. Two new releases show the potential of these developments for libraries and cultural institutions. The first is the new Media Search on Wikimedia Commons, by the Structured Data Across Wikimedia team, and the second is a proof-of-concept Image Suggestion API, by the Platform Engineering team.
Searching across languages
Media Search (or Special:MediaSearch) is an image-focused interface that makes it easier to find what you’re looking for on Wikimedia Commons. Most importantly, the search results are language agnostic. Given a search term like "zonnevlek" (Dutch for “sunspot”), Media Search won’t just return the one file on Commons that uses that term, it will search Wikidata for relevant entities and then find all files with that term and any of its aliases or translations. For the “zonnevlek” example, the number of images returned increased from one file to more than six hundred files. Media Search will make the millions of images contributed by libraries and cultural institutions much more accessible to a broad global audience.
You can try the new search here. It became the default search landing page for anonymous users on 1 April 2021, and for all users in May 2021.
To increase the search relevance of your files, you should include a descriptive title and detailed description, use the relevant Commons categories, and add depicts statements and a caption as Structured Data.
The Image Suggestion API is a service that will generate a list of unillustrated articles for any language version of Wikipedia, and then suggest up to 10 images for placement on those articles. The API will be powering a planned ‘add an image’ structured task for newcomers to Wikipedia but could also be used to drive image reuse campaigns, such as Wikipedia Pages Wanting Photos.
Right now, the API is only a proof of concept and is still being developed. You can try it at API Documentation and learn more on the MediaWiki page. If you can imagine using this API in your work with images, share your ideas on the Talk page.
The API uses algorithms that simply aggregate existing information from Wikidata and Commons, drawing on connections already made by experienced contributors. There are four main ways that it suggests matches to unillustrated articles:
Look at the Wikidata item for the article. If it has an image (P18), choose that image.
Look at the Wikidata item for the article. If it has a Commons category associated (P373), choose an image from the category.
Look at the articles about the same topic in other language Wikipedias. Choose a lead image from those articles.
Search MediaSearch for the title of the article. If an image ranks high enough in the results, choose that image.
To make your files available to the Image Suggestion API, you should use the relevant Commons categories and add depicts statements as Structured Data.
Learn more about the benefits of using Structured Data on Commons by reviewing the updated documentation and joining our April office hours on Monday, 26 April, 3.30-4.30pm UTC, and on Tuesday, 27 April, 11.00-12.00 UTC.
This was also published in the WMF This month in GLAM newsletter. You can read the rest of it here.
Hope to see you all during the April office hours.
Best,
