Understanding the Impact of Covid-19 on Mobility Behavior of Public Transport Passengers through Data Extraction

Abstract

The COVID-19 pandemic has severely impacted the mobility of people worldwide, in particular the public transport sector. For example, in Chile's capital city of Santiago, during the second week of March 2020, there was a decrease in the number of trips taken using public transport modes of 55\% and 45\% for metro and bus, respectively \cite{Astroza2020Mobility}. The isolation and work-from-home mandates that caused these decreases were commonplace worldwide and are still present in 2022 due to the continued waves of COVID-19 infections. The pandemic caused a change in people's commuting habits, either due to necessity or a feeling of safety. If or when the pandemic subsides, how many of these citizens will return to their previous commuting habits is unknown. The planning of public transport networks depends on estimated demand. The changes in the behavior of the citizens bring new uncertainty to these estimations and make the planning process much less accurate. Verifying if this trend will be hard to reverse is vital for everyone, as collective transport modes in urban centers are one of the most effective ways to prevent excess carbon emissions \cite{Tirachini2020Covid}. This dissertation aims to provide a better understanding of the mobility behaviors of passengers through analysis of data from the Automated Fare Collection (AFC) system of the Porto Metropolitan Area during different stages of the pandemic and pre-pandemic operations. The treatment of the raw data aims to create relevant data sets for the goal. Machine Learning will be used to analyze the data sets to extrapolate relevant information, possibly revealing trends in behavior. The results of this analysis can ideally be used for the planning and resource allocation of public transport in the area.