The Functional Clustering of Daily Air Quality Indices in the First Month of Previous and Current Year of COVID-19 Prevalence in Tehran

The Functional Clustering of Daily Air Quality Indices in the First Month of Previous and Current Year of COVID-19 Prevalence in Tehran

The COVID-19 impacts air quality indices during the lockdown by restricted policies for transformations, working hours and etc. Meanwhile, the different campaigns such as broadcasting advertising, social distancing and etc. help to inform people to control their behaviors. In this study, we focused on the daily air-quality indices before and in the early stage of national lockdowns that means in the first month of COVID-19 in Tehran. We compare the time series of this period with the exact previous year period with two functional clustering methods. The first one is based on the functional principal scores obtained by EM clustering, called EM Cluster and the second one is k-Centers Functional Clustering (kCFC). We study the clustering results with two indices: 1) the correct percentage of clustering and 2) adjusted Rand Index. The correct percentage of clustering for NO2 is the highest among others, between 80 and 90% and it has decreasing pattern after 20 days. Finally, we conclude that the air pollution of Tehran like other important cities in the world reduced due to the social campaign and national lock-down.