Author: Dr Megan Gooch is Head of the Centre for Digital Scholarship at the Bodleian Libraries
What do coins, ethnographic collectors, a medieval manuscript, religious texts and AI have in common? They all made an appearance at a Hackathon run by the Bodleian Libraries’ Centre for Digital Scholarship team in March 2026.
A hackathon, in case you’re new to the concept, is a collaborative (and often competitive) event using data to build a technology prototype.
Why did we do a hackathon?
Quite honestly, we’ve had data envy for a while now. Some of our brilliant library and museum colleagues in places like the US Library of Congress Labs, National Library of Scotland’s Data Foundry and the Netherlands’ Rijksmuseum Data Services have been showcasing their collections data for years. We were also inspired by the KU Leuven Libraries approach to cultural data hackathons. This approach, called Collections as Data, has been on our To Do list, but an opportunity to meet colleagues from the State Library of Berlin’s Stabi Lab as part of the Oxford-Berlin Research Partnership kickstarted our collaborative Collections as Data journey.
At the Stabi Lab they already had some data collections online and were keen to find out more about how people might use these data. At the Bodleian we wanted to know how people wanted to use data before we invested in any infrastructure or data clean-up projects.
The result was a year-long research project in which we used two hackathons – one in Berlin in October 2025 and one in Oxford in March 2026 – to see what people did with some Oxford and Berlin cultural heritage data sets. Read the results of our Berlin edition here.
Image: Attendees at the Code the Collection hackathon in the Centre for Digital Scholarship (Weston Library). Photo credit: Nick Cistone, Photographer, © Bodleian Libraries, University of Oxford.
How did people use our data?
We welcomed teams with mixed technical abilities, including MSc in Digital Scholarship students, programmers, humanists and social scientists. We added some of our own technical and digital collections specialists to the mix in case anyone needed any help.
We provided a range of datasets from the Prussian Cultural Heritage Foundation (including the Berlin State library and museums in Berlin), and Oxford’s Gardens, Libraries and Museums.
One team created a ‘vending machine’ in which you fed a coin, and it would play you a musical instrument from the same year as your coin was made. As a lifelong numismatist (coin specialist), I loved the creativity of this one. Another team used data from the Ethnological Museum in Berlin to map collectors and collecting hotspots, which revealed the obvious (German cities were popular), but also uncovered some surprising connections like collecting activity in Brazil, West Africa and Chile.
The third team came with a research question about analysing religious and non-religious texts and used a range of textual data and software Stata to demonstrate the changing nature of languages and theological texts. The fourth group took one manuscript, the medieval Piers Plowman at the Bodleian, and created an interface which both enabled those studying Middle English texts to interrogate different dialect words and had handy features for Middle English learners to look up words. Our final group was one of our software engineers who unleashed his creativity by using AI to create an interactive discovery interface for Berlin collections.
What did we learn?
For me, the most amazing thing about what our hackers created was what they did with data we already had openly available on the internet. I mean, yes, we pointed people to the data and collated a list of datasets. But these amazing games, analytical tools, and entire discovery platforms didn’t require us to spend years creating new digital infrastructure or enhancing our metadata.
Could we have better infrastructures for computational access? Well yes, there’s always room for improvement and technology is changing so quickly. Could we have better data? Also yes, but no data will ever be perfect, so it’s a real insight to know that researchers, students and programmers can deal with the messiness of real cultural heritage data.
My final learning is that it takes a heck of a lot of work to organise a hackathon, and the team did a brilliant job of making it look this easy. We’re hoping to publish more blog posts on our hackathon teams, so sign up to our mailing list to stay in the loop.
What: