Fig. 4: Validation of the hotspot-detection method on the Beijing dataset.
From: Detecting urban PM2.5 hotspots with mobile sensing and Gaussian process regression
Panels (a, c) depict the spatial PM2.5 distribution from ref. Wang et al.43 and panels (b, d) depict the hotspot scores calculated by our method. The colours correspond to the intensity for every 30 × 30 m grid tile. The hotspot scores generally match the spatial pollution pattern in the city, although the sharp boundaries between high and low pollution are sometimes smoothed over.
