Urban regeneration increasingly depends on understanding how mobility and accessibility shape who can reach jobs, services, and public space.
The researchers explain that mobility is increasingly intertwined with modernising infrastructure, closely linked to economic revitalisation, health, and social equity by connecting people to opportunities.
As a result, there is a need to enhance accessibility and foster social inclusion, particularly for marginalised communities and why planning truly walkable and equitable cities requires accurate, city-wide pedestrian activity estimates integrated with regeneration strategies.
Led by Mohamed Shamroukh, a Doctoral Researcher at the School of Architecture, Building and Civil Engineering at Loughborough University, and supervised by Dr Asya Natapov and Dr Taimaz Larimian, the researchers introduced the Geospatial Pedestrian Activity Engine (GPAE), a framework integrating spatial analytics with machine learning to predict pedestrian activity across entire urban areas beyond observed locations.
The system, originally developed as part of the Loughborough Town Deal, was tested in Loughborough using data from just 29 monitoring locations, combining information from WiFi sensors, video cameras, and details about the local environment. It then used machine learning to estimate pedestrian activity across the whole town.
Published in Travel Behaviour and Society, the model was found to be highly accurate and found that pedestrian activity is influenced not only by transport links, but also by the presence of shops, services, community facilities, and nearby housing.
The findings showed that a relatively small amount of pedestrian sensor data, combined with AI and geospatial information, can accurately map walking activity across an entire city and provide planners with a practical framework for targeting urban regeneration and active travel investment.
By combining predicted footfall data with information about existing infrastructure, the tool helps planners identify areas that are:
- Overburdened: lots of people but not enough infrastructure
- Underutilised: infrastructure exists but few people use it
- Neglected: fewer people and limited investment
- Effectively supported: infrastructure matches demand
The research also demonstrated how existing pedestrian data can be transformed into decision-ready evidence for active travel schemes; walking infrastructure upgrades; public realm improvements; and sustainable mobility planning, rather than relying solely on reactive observations or limited manual counts.
It is hoped that these findings will help local authorities make better decisions about where to invest in walking routes, public spaces, and regeneration projects, ensuring funding is directed to the places that need it most rather than relying only on limited observations.
Mohamed Shamroukh said of the research: “Walking is one of the most important ways people experience and move around a town, yet we often have very limited information about where it happens. Our research shows that AI can turn small amounts of data into a much bigger picture of pedestrian activity, helping planners see where support is needed most. Ultimately, we hope it leads to more walkable, inclusive and sustainable places for everyone.”
The full research is available here.