Detecting Urban PM$_{2.5}$ Hotspots with Mobile Sensing and Gaussian Process Regression

Low-cost mobile sensors can be used to collect PM2.5 concentration data throughout an entire city. However, identifying air pollution hotspots from the data is challenging due to the uneven spatial sampling, temporal variations in the background air quality, and the dynamism of urban air pollution sources. This study proposes a method to identify urban PM2.5 hotspots that addresses these challenges, involving four steps: (1) equip citizen scientists with mobile PM2.5 sensors while they travel; (2) normalise the raw data to remove the influence of background ambient pollution levels; (3) fit a Gaussian process regression model to the normalised values; (4) calculate spatially explicit ‘hotspot scores’ using the probabilistic framework of Gaussian processes, which summarise the relative pollution levels throughout the city. We apply our method to create the first ever map of PM2.5 pollution in Kigali, Rwanda, at a 200m resolution, where we uncover several pollution hotspots. We also evaluate our method using simulated mobile sensing data for Beijing, China, where we show that the hotspot scores capture the ground-truth spatial PM2.5 distribution. Thanks to the use of open-source software and low-cost sensors, our method can be re-applied globally to help fill the gap in urban air quality information.