LAUSR

Abstract This study examines the effect of weather on bikeshare use. We employ data from forty Public Bicycle Sharing Programs located in forty cities (16 countries) across five different climate zones, spanning tropical to boreal climates. Our curated dataset is longitudinal and consists of nearly 100 million cycling trips. Key findings include: (a) the most significant variable, particularly on weekdays, is the time of day, followed by precipitation; (b) in most cities, usage increases on weekdays and weekends up to a point around 27 to 28 °C, before declining; (c) usage by hour usually follows a bimodal or trimodal daily pattern on weekdays, except for schemes which are too small to serve a commuter function (weekend and weekday usage is similar in small schemes); (d) weekend usage peaks at around 2 to 3 pm in most schemes, except those in hotter climates where the peak is around 5 pm; (e) precipitation negatively affects female ridership more than male ridership; and, (f) a changing climate is likely to affect cycling by boosting ridership in cold climates and lowering ridership in warm climates, but the effects will likely be small. In the spirit of reproducibility, all data and R code are publicly available.