Air quality analysis in critical zones of Mexico using unsupervised machine learning

Air pollution continues to be a significant risk to public health. It is an important issue that affects everyone and should be studied to take actions that help mitigate the risk. This research’s main objective is to explore air quality from 2010 to 2018 in metropolitan zones of Guadalajara, Monterrey, and the Valley of Mexico. Air quality is analyzed based on the concentrations of criteria pollutants: carbon monoxide (CO), nitrogen dioxide (NO2), ozone (O3), particulate matter less than or equal to 2.5 micrometers (PM2.5), particulate matter less than or equal to 10 micrometers (PM10), and sulfur dioxide (SO2). In this paper, six stages are proposed:1) data collection; 2) data cleansing; 3) descriptive analysis. 4) design of cluster analysis models and main components; 5) application of machine learning models; 6) interpretation and analysis of results. This study has concluded that higher mean hourly concentrations for pollutants, CO, NO2, O3, PM2.5, and SO2 were found in Guadalajara Metropolitan Zone. In contrast, the highest mean hourly concentration of PM10 was found in the Monterrey Metropolitan Zone. Furthermore, the pollutant with the highest effect in the three critical zones was PM2.5, presenting high concentrations with hazardous air quality levels. Finally, the area with the worst air quality was Guadalajara Metropolitan Zone.