Big data analytics to classify earthwork-related locations: A Chengdu study

Air pollution has significantly intensified, leading to severe health consequences worldwide. Earthwork-related locations (ERLs) constitute significant sources of urban dust pollution. The effective management of ERLs has long posed challenges for governmental and environmental agencies, primarily due to their classification under different regulatory authorities, information barriers, delays in data updating, and a lack of dust suppression measures for various sources of dust pollution. To address these challenges, we classified urban dust pollution sources using dump truck trajectory, urban point of interest (POI), and land cover data. We compared several prediction models and investigated the relationship between features and dust pollution sources using real data. The results demonstrate that high-accuracy classification can be achieved with a limited number of features. This method was successfully implemented in the system called Alpha MAPS in Chengdu to provide decision support for urban pollution control.