AI-driven map could link UK air quality to health

Developed by the London School of Hygiene & Tropical Medicine (LSHTM), the research combined data from Earth observation satellites with readings from ground-based pollution monitors, as well as other information such as population density, road density and the location of airports. Machine learning algorithms were then applied to these datasets, producing estimates of ground-level concentration of fine particulate matter (less than 2.5 microns, PM2.5) for the whole of Great Britain from 2008-2018.

IMechE calls for new clean air act to tackle ill-health from air pollution

The team mapped the results onto the island of Great Britain using a 1 x 1km grid and cross-referenced the estimates using historical data from ground-based monitoring stations, with the modelling found to be ‘closely aligned’ to actual readings. South-East England was revealed as the most polluted region, but air across the whole island was found to be improving over the decade-long period. It is hoped that the granular picture of air pollution can now be used by epidemiologists to study the links between air quality and healthcare problems.