Exploring the Effects of Teleworking on Travel Patterns: A Simulation-Based Study with MATSim

Abstract

This study assesses the impact on travel patterns of teleworking in Quebec City, Canada, using the MATSim/Eqasim model. It compares scenarios where (a): teleworkers stay home without altering daily activities or (b): engage in new activities, like driving children to school or additional shopping and leisure trips. The analysis uses origin-destination surveys and demographic data to evaluate modal shares, travel distances, and congestion changes. In previous practice, teleworking activity is assigned to random agents to generate scenarios, which may not reflect real-life conditions. This study fills that gap by predicting individuals’ probability of teleworking based on socio-demographics, built environment, and assigns teleworking activities to those who are more likely to telework. The analysis is based on an ensemble learning method, Extreme Gradient Boosting. Findings show that teleworking significantly reduces car usage and overall travel distances while increasing short walk trips, leading to more sustainable urban mobility and reducing congestion.