Unlocking Community Well-being with Administrative Data and AI
Rafael Carrascosa Marzo
What makes a community thrive? It’s a deceptively simple question, but one that has challenged policymakers, researchers, and local leaders for decades. Traditionally, the answers have come from surveys: people telling us how happy they feel, how connected they are, or how satisfied they are with their lives. These are valuable insights, but surveys are expensive, slow to run, and don’t always give us the whole picture.
A new report, Using administrative data and artificial intelligence to understand community well-being, explores a different route: turning to the everyday data that local authorities already collect. Think library memberships, recycling records, parking permits, or even noise complaints. This “administrative data” is everywhere and widely used in the private sector, but until now it’s been under-used in the public realm.
The research team, spanning UCL’s Department of Psychology and Institute for Global Prosperity, Edge Hill University, and Emmanuel College, Cambridge, wanted to see whether combining these existing data sources with AI could provide sharper insights into how communities are really doing. Funded by the Nuffield Foundation and the British Academy as part of their Understanding Communities programme, the project partnered with Camden Council to test this idea in practice.
Why now?
The timing couldn’t be more relevant. Local authorities are under severe financial pressure, and the demand for better, faster, and more cost-effective ways of measuring well-being has never been greater. At the same time, national conversations about data infrastructure and AI in government are gathering pace. With inequalities widening and economic growth no longer translating straightforwardly into improved living standards, finding new ways to understand and support community well-being is no longer a “nice to have”, it’s essential.
From surveys to dashboards
The project developed a proof-of-concept Community Well-being Index Dashboard based on the ‘Good Life Euston Model’ as a starting point for understanding community well-being. By running administrative data through an AI model, the researchers were able to generate well-being scores for every local authority in England. What’s striking is how well these scores correlate with productivity, in fact, they perform better than traditional subjective well-being measures.
This matters because if a place is struggling economically, policymakers could use this data to pinpoint which aspects of community well-being (housing, education, social connection, or others) are falling behind. That opens the door to more targeted interventions, better resource allocation, and ultimately, a stronger link between well-being and inclusive local development.
But… it’s complicated
Of course, the work also exposed some of the barriers. The report highlights four big challenges:
Infrastructure isn’t there yet. Local authorities simply don’t have the systems in place to fully harness administrative data. National reform, from harmonising data structures to training staff, is needed to unlock its potential.
Data governance needs a rethink. Communities should have a say in how their data is used, who can access it, and what protections are in place. Without legal frameworks that build in consent, privacy, and fairness, public trust won’t follow.
Trust is fragile. While people give up data to private companies every day (often in exchange for free services), they’re more hesitant about government use of their information. That means AI in the public sector must be transparent, explainable, and co-designed with communities.
Well-being is complex. There’s no single measure that works everywhere. More research and more test cases are needed to understand how well-being interacts with wider social and economic factors.
Where next?
Despite the challenges, the report is optimistic about what’s possible. Administrative data and AI could transform how councils understand their communities, not only on well-being, but also on health inequalities, poverty, and other pressing social issues. But to get there, we need investment in infrastructure, stronger governance, and a willingness to experiment.
This research is an invitation: to rethink how we use the data we already collect, to build trust through transparency, and to explore how new technologies can support a more inclusive and sustainable future.
Community well-being isn’t just about how people say they’re doing, it’s about the patterns and signals we can observe in daily life. When combined with AI, those signals may offer powerful new ways of understanding, and ultimately improving, the places we live.
Rafael Carrascosa Marzo is Project Manager for PROCOL UK.