Hack the data, shape the future - introducing the datathon!
Owen Davies, our Head of Strategic Data Programmes, took part in the 2025 Datathon hosted by ADR Wales. In this blog, he shares what a datathon is, how teams tackled a challenge around young people’s transition to adulthood, and why events like this matter for social care.
This year Wales hosted the Administrative Data Research conference. The main event was held in the Welsh College of Music and Drama, but several fringe events were held across Cardiff over a course of a week in September.
This is when I was invited to take part in something called a “datathon”.
“What’s a datathon?” I can hear you say. Well, let me explain….
What is a datathon?
The datathon is a portmanteau (two words squashed together to make a new one) of data and marathon.
A datathon is an event much like a hackathon where people come together and form teams to build a prototype solution to solve a challenge, typically involving software or technology.
Many hackathons offer prizes, judging panels and presentations at the end of a two-day session. The difference between a datathon and a normal hackathon is there’s an expectation that the challenge is met by using both technology and linked data analysis.
The datathon challenge question
For this event, Social Care Wales were invited as the “challenge owner”. The challenge owner is responsible for writing the question that the teams would try to “hack” in the two days (or about 10 hours) they were allowed.
Our question was centred towards the transition of young people into adulthood. We asked whether there were differences in services young people could expect, particularly in relation to mental health services, when they turned 18.
Broad questions like this are really good for these sessions as there are lots of different types of analysis and different techniques that can be used.
The challenge
Four teams of six researchers with a range of backgrounds (data scientists, social researchers, computer scientists) were put together to try and solve the challenge. Each team had access to several synthetic datasets from the SAIL databank.
A synthetic dataset is one that looks exactly like the real thing but has made up data. This makes it easier for people to have access to data without needing to go through lots of information governance and data ethics processes.
The synthetic datasets ranged from GP data to PEDW (Patient Episode Database in Wales which is used with hospital in-patients), to the Children Looked After and Children Receiving Care and Support Census, and much more.
The teams could link datasets together to start gathering more insights.
A team of experts and mentors were available throughout the challenge to help people with any technical questions they had.
Datathon: the judging panel and process
Now, the process of the datathon isn’t about the results. The data is synthetic, so any of the outputs are meaningless. The winning team is chosen using several criteria areas, such as:
- creativity, is it a novel or an original idea?
- technical ability, is it ambitious and functional?
- impact, would this give us new insight or understanding?
- teamwork, has the team worked well together?
- quality, is the work of high quality?
- presentation, have they told a compelling story?
The judging panel were made up of researchers from Wales, London and Singapore, and our very own Lisa Trigg was one of the panel.
Judging is often a difficult task. Teams work really hard on challenges and lots of really interesting and high quality work gets done.
The winning team’s outcomes
The winning team (Team 4) looked at the impact of social determinants on access to mental health services across ages.
They managed to incorporate all five of the synthetic datasets, had a good team vibe and told a compelling story about why they’d chosen this path in their final presentation.
Well done to Team 4, and well done to all those who took part!
Final thoughts
If you see an event like this advertised near you, consider taking part!
Datathons aren’t just for data and tech nerds, but social researchers and practitioners too. You’ll be tired by the end I can assure you, but you’ll have met new people, figured some things out and learned something new.
Thank you to ADR Wales for inviting me to be a part of the 2025 Datathon - it was great to be asked to take part.
Now, when’s the next one….