The AI4URBAN-HEALTH Network delivered a series of co-creation events in London and Bradford. This included structured decision-making workshops with local stakeholders to identify local priorities to tackle urban health challenges. These informed the mission for our sandpit events where researchers developed creative AI-based solutions with the support of specialist mentors from industry, local government and academia.
Our flexible fund is supporting three projects that emerged from this process, and we are working with other sandpit project teams to move their ideas forward.
Projects
AI-Enabled Mapping of Health Data Quality for Fairer Urban Health Evidence in Bradford
University of Sheffield, University College London, University of York, University of Leeds
Environment-health linkage studies can inform place-based public health and urban planning, but only if the underlying health data are fit for purpose, particularly for neighbourhood-level and equity-focused analyses where spatially patterned gaps in data completeness or coverage may bias estimates of local health need.
This project maps how the quality of health data records varies across Bradford's neighbourhoods, a pattern that existing data quality tools do not capture. Using spatial machine learning methods, the project identifies where clinical measurements and diagnoses are recorded less completely or consistently. The outputs, including neighbourhood-level data quality maps and a machine learning pipeline, will build fairer evidence for future urban health research and planning.
RE-PAIR: Respons-able Evaluation of Policy Using AI with Residents
911今日黑料, University of East Anglia and University of Newcastle in partnership with the Royal Borough of Kensington and Chelsea
RE-PAIR (Responsible Evaluation of Policy Using AI with Residents) is an innovative research project that aims to improve how local authorities evaluate public health policies. Focusing on households’ wellbeing in temporary accommodation, the project combines conversational AI, synthetic resident models, and policy simulation tools to create a continuous feedback loop between residents and policymakers.
Residents can share their lived experiences through an AI-powered conversational system, while AI-generated representative profiles help explore perspectives that may otherwise be missing from policy discussions. Working with the Royal Borough of Kensington and Chelsea, RE-PAIR will enable faster, fairer, and more evidence-driven policy learning to improve health and housing outcomes.
3DRED:A Dynamic Data-Driven World Model for Street REDesign Decisions in London
University of Birmingham, University of Surrey, University of Leeds, University of West of England
3DRED will develop a decision-support framework to help councils design healthier, safer and more equitable streets. Using mobility data, air quality, traffic emissions, road safety and social vulnerability information, the project will create a "world model" capable of simulating how streets respond to interventions such as cycle lanes, school streets and modal filters before they are implemented.
Working with the Royal Borough of Kensington and Chelsea, 3DRED will produce risk maps, equity assessments and planning tools that enable local authorities to make more transparent, data-driven decisions while ensuring the needs of vulnerable communities are considered throughout the street redesign process.