Air quality vote distribution is strongly correlated to the Air Clean/Dusty votes, and both of these perception votes
are correlated with the PM count. Overall as the concentration of
PM increases the number of ‘clean air’ and ‘good air quality’ votes
decreases, whereas the number of ‘dusty air’ and ‘poor air quality’
votes increases. Through the exposure–response relationships between
the various perception votes and PM, it was possible to predict
air cleanness using the PM count. The model performed fairly well for