Growing visual knowledge on the Wikimedia projects: the experiences of the Image Description Month

Translate this post
A banana is on a table and the image of a banana is on the screen of a laptop, both are framed and labeled as "banana"
Bounding boxes are drawn around both a physical and an on-screen banana as a metaphor for computer vision (Max Gruber, CC BY 4.0, via Wikimedia Commons)

Today, our lives are fully connected to the digital environment, through social media, mobile apps, and for those who participate, the complex ecosystem of the Wikimedia projects. In these online surroundings, images are increasingly present and becoming a predominant way of communicating. On English Wikipedia, for example, there are more than five million distinct images and research that investigated readers’ interactions found that images on Wikipedia “drive 10 times the engagement that citations do.” (See Rama, D., Piccardi, T., Redi, M. et al. (2022) A large scale study of reader interactions with images on Wikipedia – EPJ Data Sci. 11, 1).

For this reason, visual and media literacy is more important than ever and, in Wikimedia projects, visual knowledge is both its own form of knowledge, and a necessary support for encyclopedic knowledge. To ensure visual knowledge is available to everyone, it must be made findable and accessible through alt text and additional description and data.
With all of this in mind, in October 2023, the Wikimedia Foundation’s Culture and Heritage team supported a second edition of the “Image Description” initiative, extending it from a week to a month. During this Image Description Month, we had a total of nine events organized by seven different communities and partners spread across four different continents (North America, South America, Europe, and Asia): Digital Public Library of America and Wikimedians of Indiana, CIS-A2K India, Biodiversity Heritage Library and OpenRefine, Wikimedia Colombia, Wikimedia Argentina, and Wikimedistas de Uruguay.

To document the initiative and shine a light on the events organized, as well as the communities who participated in and delivered them, we have gathered first-hand reports. When available, links for recordings and information pages are also included.

Events by the Digital Public Library of America and Wikimedians of Indiana

WikiWednesday @ IUPUI and the DPLA Network Coffee Chat

The Digital Public Library of America organized two events for the IDM. WikiWednesday @ IUPUI was organized with Wikimedians of Indiana and held in person at the university library. Here’s an account from Jamie, one of the organizers of the event:

“WikiWednesdays are on the first Wednesday of every month, when the user group gets together with librarians at the University Library at Indiana University to teach Wikipedia-related topics. October’s theme was around Image Description, where we used digital items from Wikimedia Commons that had been digitized by the library and uploaded through the DPLA pipeline to talk about improving image description and “thinking outside the box” for how to use images across Wikipedia. For example, using an image of a girl’s worsted wool dress to illustrate the Worsted article. The dress comes from the Conner Prairie digital collection in Indiana Memory.” – Jamie Flood, Wikimedians of Indiana User Group member, DPLA Wikimedia working group member, and Senior Wikipedian and Outreach Specialist, National Agricultural Library.

DPLA’s second event, DPLA Coffee Chat: A Wikimedia Case Study with Ohio Digital Network and Columbus Metropolitan Library, was organized by its Wikimedia working group. In this online event, participants shared their experiences contributing over 350,000 images from the Columbus Metropolitan Library’s digital collections using the Digital Public Library of America’s pipeline to Wikimedia Commons.

After the presentation, there was a group discussion and interactive editing session to add images, captions, and Depicts statements to DPLA files on Wikimedia Commons. This part was not recorded, but participants were able to edit several files, ask questions, and request help from members of the Wikimedia working group. The recording of the first part of the event is available on YouTube:

Recording of the DPLA Coffee Chat organized by the Wikimedia Working Group

Events by Wikimedia Colombia

Abrir la Cultura y el Patrimonio: seminars 3 and 4 

“Abrir la Cultura y el Patrimonio” (in English, “Opening up Culture and Heritage”) was a series of four seminars to provide an overview of the possibilities for collaboration between cultural institutions in Colombia and the Wikimedia movement. For Image Description Month, Wikimedia Colombia contributed their third and fourth seminars: one about how to open collections and digitization projects, and another about how to make collections visible through metadata.

Here’s a translated first-hand report from Paula, the main organizer:

“The seminars arose from the need to create spaces for meetings, learnings, and reflections about the possibilities of cultural action in Colombia. Each seminar followed the sequence Problem → Wikimedia Tool→ Solution→ Impact, allowing participants to find commonalities or points of connection in each presentation. Over 60 people participated and 32 were certified due to their dedication and interest. This experience has also created momentum in cultural institutions, generating interest and anticipation for a future edition and potential collaborations and support from Wikimedia Colombia.

Recording of the third session with Wikimedia Argentina

For the 3rd session, Angie Cervellera and Mauricio Genta from Wikimedia Argentina shared their experience with cultural institutions. The session explored the concept of open collections and activities that extend beyond Open Access. Mauricio then delved into the challenges Wikimedia Argentina faced when the project was initially proposed and detailed each phase that a Wikimedia and digitization project needs to consider: planning, preparation, capture, and treatment. Mauricio concluded by delving into the legal frameworks that must be considered when uploading collections to Commons. The main objective of this session was to raise questions about the link between digitization and dissemination of heritage images and it attracted 27 managers from cultural institutions across the country.

Recording of the fourth session with Silvia Gutierrez

For the 4th session, Silvia Gutierrez created a series of interactive exercises to promote reflections on collective knowledge production. Silvia delved into an explanation of Wikidata, emphasizing how it’s collectively constructed. After the presentation, participants were invited to explore the Wikidata Query service. With Silvia’s guidance, they constructed queries with the objective of discovering essential information about museums, libraries, and cultural institutions in Colombia. This exercise allowed the participants to identify content gaps and understand how they could engage with them on the Wikimedia projects.” – Paula Durán, Culture Program Officer, Wikimedia Colombia.

Events by CIS-A2K India

Image Description Month in India (SDC and CIS-A2K sessions)

CIS-A2K India organized two events. The first was an online event (recording) introducing Structured Data on Commons, where participants edited and added data to images. 

The second was a month-long invitation for each Indian language community to describe images on their local Wikipedia, plus Wikidata and Wikimedia Commons (report).

Here’s a first-hand report from Nitesh, the main organizer: 

“Regarding the recent Image Description Month campaign, I wish to highlight its significance in enhancing the accessibility and usability of visual content on Wikimedia Commons across all languages. In collaboration with CIS-A2K and the Image Description Month organizing team, we aimed to promote this campaign within Indian communities. Our efforts resulted in the participation of 21 Indian Wikimedians in the content creation or development. There was also an engaging advanced structured data training session conducted led by Giovanna Fontelle (WMF), which garnered positive feedback and valuable suggestions from some attendees.

I’m pleased to report the significant contributions made by editors during the campaign:

  • 243 captions were meticulously edited
  • 1 caption received significant modifications
  • 3007 depictions were thoughtfully added
  • 54 depictions were updated for accuracy
  • 215 new uploads enriched Wikimedia Commons
  • 607 categories were thoughtfully curated to enhance discoverability

Buoyed by the overwhelmingly positive response from the Indian community, I’m eager to sustain this collaboration for the future and expand upon the achievements. If there are plans underway to organize a similar event this year, I would make a step forward with a renewed focus on innovation and inclusivity. Indeed, campaigns of this nature not only provide a platform for volunteers to collaborate but also ignite a spirit of collective action on both national and global scales. As the journey continues, the possibilities for growth and impact are boundless, fueled by a shared commitment to knowledge accessibility and community empowerment.” – Nitesh Gill, Program Officer, Centre for Internet and Society.

Event by the Biodiversity Heritage Library

The Biodiversity Heritage Library (BHL) organized an event titled Transforming Biodiversity Heritage Library Images – Data Modeling with OpenRefine (recording). This event’s objective was to develop a Structured Sata on Commons model for BHL. The event was supported by Sandra Fauconnier, as part of her work with OpenRefine. 

Here’s a first-hand report from JJ, one of the event’s organizers:

The BHL-Wiki working group has been scaling up the BHL to Structured Data on Commons pipeline to improve image search across the vast biodiversity image collection

“In collaboration with Wikimedia Foundation’s Giovanna Fontenelle, Wikimedian Sandra Fauconnier, and BHL Data Manager JJ Dearborn, the workshop united BHL Staff, Wikimedians, Flickr staff, and biodiversity image enthusiasts. The primary focus was on mapping and loading BHL Flickr image data to Structured Data on Commons (SDC), a key initiative by Wikimedia Commons. This project, identified as the second top-voted priority in the BHL Wikimedia Whitepaper, “Unifying Biodiversity Knowledge to Support Life on a Sustainable Planet,” made significant progress, covering 60% of the mapping work for 27 out of 45 fields. Sandra provided an exciting preview of the Wiki Commons extension for OpenRefine.

Looking ahead, there was much enthusiasm to sustain momentum from the 16 participants who attended, prompting plans for Part 2 of the workshop. The aim is to complete the mapping work collectively while delving deeper into OpenRefine, exploring its capability to load image data into the Structured Data Commons ecosystem. This effort not only establishes a SPARQL endpoint for BHL’s Image Collection but also unlocks potential data harvesting back into the BHL data ecosystem. The ultimate goal is to enhance the accessibility and utilization of public domain images for users across Wikimedia Commons, Flickr, and BHL. Stay tuned for updates on Part 2.” – JJ Dearborn, Data Manager, Biodiversity Heritage Library.

Event by Wikimedia Argentina

Poster of the “Experiencias culturales accesibles: del territorio a la pantalla” event by Wikimedia Argentina (Angie Cervellera (WMAR), CC BY-SA 4.0, via Wikimedia Commons)

The event with Wikimedia Argentina was called “Experiencias culturales accesibles: del territorio a la pantalla” (in English, “Accessible cultural experiences: from the territory to the screen”). The recording is available.

Here’s a translated first-hand report from Angie, the main event organizer:

“Wikimedia Argentina’s Culture and Open Knowledge program held a virtual meeting addressing accessibility as a transversal perspective in cultural spaces. As part of the Image Description Month, the session titled “Accessible cultural experiences: from the territory to the screen” aimed not only to raise awareness about the issue but also to encourage participants to take action on the matter, in this case, by learning how to edit Wikipedia’s alternative texts. There were 30 participants, mostly Spanish speakers, and they improved the alt texts in 5 articles.

The presentation was given by Gladys Benítez, scriptwriter, narrator of audio descriptions for series and films, and member of TNC (the accessible program of the Teatro Nacional Cervantes). In it, different accessible resources for blind or visually impaired people were covered.

After, a space for practice and exchange was set up so that participants could collectively produce alt texts for images on Wikipedia. We divided ourselves into groups of five and worked on different local articles with images. Each group collaboratively discussed what would be the best description for each image and the edits were made.

We believe that these types of proposals are especially interesting for all cultural workers. It’s important that they understand that accessibility is a collective commitment and that it largely depends on small actions that, when added together, lead to practices that are increasingly accessible to everyone. We emphasize, along with Gladys, that accessibility makes people feel part of a space and want to stay (or return), and we think that the same happens with Wikipedia, understood as a global community based on the voluntary participation of people. We, therefore, emphasize the importance of small actions such as editing ALT texts so that Wikipedia becomes increasingly a more friendly space, where everyone can feel welcome.” – Angie Cervellera, Culture and Open Knowledge, Wikimedia Argentina.

Recording of the “Accessible cultural experiences: from the territory to the screen” organized by Wikimedia Argentina

Event by Wikimedistas de Uruguay

Wikimedistas de Uruguay organized the month-long initiative “Click a la escuela” (in English, “Flash to the school”). Here’s the on-wiki description by Fede, one of the organizers: 

In the framework of Heritage Day 2023, Wikimedians of Uruguay proposed a photography contest to document school buildings throughout the country. The call was open from October 1 to 31 inclusive. During this period, people could upload photographs of schools and school buildings through the WikiShootMe tool, which allowed them to easily upload photographs with geographic coordinates. – Fede Colman, Wikimedia and Heritage Institutions Fellow, Wikimedistas de Uruguay.

Can you help us translate this article?

In order for this article to reach as many people as possible we would like your help. Can you translate this article to get the message out?