Understanding walking behaviour during COVID-19: Binary Logit and ANN approach

Abstract

The COVID-19 pandemic had severe impacts on society. It negatively affected many sectors, and transportation is one of those. Naturally, the walking trip behavior of individuals was also altered. This study aims to investigate the changes in walking trips of individuals with two models: Binary Logit (BL) and Artificial Neural Networks (ANN). An online survey was conducted with 387 individuals. BL model investigated if respondents' walking trips would increase during and after the pandemic. On the other hand, ANN models were developed to determine the significant factors in changes in walking behavior. Results indicate that ANN models capture more factors than BL models. Demographics and attitudes towards public transportation, taxi, and walking trips during the pandemic are found to be effective in walking behavior changes. Policies can be made to increase public transit ridership, and infrastructure for walking can be improved. Future research suggestions are given.