Uncovering heterogeneous effects of link-level street environment on e-bike and e-scooter usage

Abstract

This study investigates how the street environment influences e-bike and e-scooter flows at the link level, considering their distinct spatial travel patterns. An angle-based spatial autoregressive quantile regression (SAQR) model is developed to analyze fine-scale street environments in Washington, D.C. We observe distinct spatial travel patterns between e-bikes and e-scooters, and e-scooter usage is more concentrated in city centers. Link design and network design have stronger impacts on the usage of e-bike and e-scooter than land use features. However, land use features are more likely to affect the flow of these two modes differently. Specifically, streets with dedicated bike lanes, traffic signals, wider width, higher betweenness centrality, and a higher proportion of entertainment and office land tend to attract more e-bike and e-scooter trips. In addition, bike-friendly facilities, particularly buffered bike lanes, exhibit more pronounced impacts. The findings provide policy implications for nuanced street design guidelines to facilitate electric micromobility usage.