Demonstrating the power of AI in land use planning

Partnering with the Geospatial Commission to simulate the effects of planning decisions on residents’ quality of life

Last updated
Tuesday 17 Sep 2024

The UK is facing difficult decisions about how it uses its land. A growing population is putting increasing pressure on housing and transport infrastructure, but urban development needs to be balanced with protecting the environment and ensuring energy and food security. The Turing has been exploring how data science and AI can help planners and policy makers to predict and understand the impacts of their decisions, working with the Geospatial Commission (part of the UK government’s Department for Science, Innovation and Technology, DSIT) to trial some pioneering technology in Newcastle, England. 

In 2023, Turing researchers led by Dani Arribas-Bel created DemoLand – a prototype tool that uses a suite of machine learning models to simulate the effect of land use decisions on the population’s quality of life. Newcastle City Council planners fed into the design of the tool, which can be used to explore the potential impacts of different development scenarios (such as converting brownfield land into housing developments or creating new parks around the River Tyne) on four key metrics: air quality, access to jobs and green space, and house prices.

This demonstrator (available to try here) is not currently designed for deployment, but the team is developing it further by incorporating technology based on the large language models that power chatbots such as ChatGPT, which will allow users to ask the tool questions like, for example, “what is the impact of this scenario on air pollution near schools?”. 

The researchers are also investigating using satellite imagery to speed up the process of gathering data about current land use. They plan to work with more local authorities to test the tool, with the ultimate aim of seeing data science and AI become an integral part of planning decisions across the country.

“In Newcastle, we require over 1,000 new homes per year until 2030. The success of this pilot project with the Turing has encouraged us to approach city planning in a more data-driven way, not just for predicting the impact of our decisions, but also for explaining them to residents.” 

Emma Warneford, Senior Specialist in Planning, Newcastle City Council

This piece first appeared in The Alan Turing Institute’s Annual Report 2023-24 
Top image: Zhou

Authors

Research areas