A0313
Title: Studying the COVID-19 lockdown effects on Iranian traffic behavior in three calendars with functional data analysis
Authors: Mohammad Fayaz - Allameh Tabatabaei University (Iran) [presenting]
Abstract: The COVID-19 lockdown has affected many aspects of people's lives, like air pollution, economics, traffic, etc. We have collected spatio-temporal traffic datasets between provinces in Iran from March 2010 to January 2023 from more than 2500 count data stations. We have studied four indices, including total traffic, and the number of traffic offenses - speeding, unsafe distance, and Illegal overtaking in the following time periods: before, during, and after COVID-19 lockdowns. The Iranian official calendar is the Solar Hijri calendar, while the Islamic Hijri calendar and Gregorian calendar are also important because many holidays and events are occurred due to them. The time series decomposition methods that consider multiple seasonal periods are applied. The outlier detection methods from fda.usc R package is used. The forecasting scenario is without the COVID-19 lockdowns and estimates the lockdown effects. The Interval-wise testing (IWT) results for comparing before and after lockdown are presented with adjusted p-values. The spatial variability of count stations near roads is estimated with functional and spatial statistical methods. The R codes are available on the author's GitHub page for reproducible purposes.
