Riders on the Storm: Exploring Meteorological and Temporal Impacts on Shared E-Scooters (SES) in Munich, Germany

This paper analyzes the meteorological and temporal impacts on shared e-scooters (SES) over 27 months of service in Munich. The objective is to explore the factors associated with SES utilization (hourly usage counts, median ride distances, and booking durations), focusing on time-variant variables (weather, holiday, time of the year, week, and day). This study employs the negative binomial (NB) and Consul's generalized Poisson (GP-1) regressions for modeling SES hourly demand. The Poisson regression is used for hourly medians of SES ride distances and booking durations. Random forest models evaluate the relative importance of meteorological and temporal variables for SES usage. In Munich, the popularity of SES grew over time. The peak booking numbers were on Fridays, Saturdays, and afternoons. Longer rides were on the weekends and holidays than on working days. The most extended trips were around midnight, posing the issue of riders' visibility. The COVID-19 lockdown negatively impacted SES bookings. Compared to winter, more and longer rides were between July and November. The weather impacted e-scooter usage with fewer bookings and shorter rides when raining and humid and more and longer trips when warm. Negative weather impacts for e-scooters may be partially due to a reduction in recreational use as weather discourages many outside activities.