Scootin’ in the rain: Does weather affect micromobility?

Abstract

The usage of shared e-scooters, dockless e-bikes, and docked bicycles are correlated with weather conditions to examine the relative impact on each mode, specifically number of trips taken, their duration, and distance. Data is obtained from the City of Austin data portal. Rain, temperature and wind conditions are obtained from NOAA and a variety of analysis methods are applied, specifically Prais-Winsten and Negative Binomial regressions as well as a Random Forest model to examine the full suite of weather variables and to avoid some of the distributional issues in the trip models. In addition, controls are included for holidays, days of the week, and special events in Austin (such as the SXSW festival); all are found to be critical control variables, with SXSW associated with large increases in trips. Results suggest that docked bicycle and e-bike usage is more sensitive to adverse weather conditions than e-scooter trips, though all are reduced in colder, rainier, and windier conditions, as well as extreme heat and high relative humidity.