Cities need to decide where shared scooters fit within their transport networks. A single pilot offers a limited view: observed demand may reflect the weather, the stage of service development or the characteristics of the area. Comparing places and periods helps separate recurring patterns from local circumstances.
Methods
We analysed approximately nine million trips from Austin, Calgary, Chicago, Louisville and Minneapolis. Trip records were joined to open weather, census, land-use and transport-infrastructure data, GTFS public-transport information and a local transit-accessibility index. Exploratory comparisons described hourly and daily demand, trip duration, distance and speed. Zero-inflated negative binomial models then accounted for the large number of place-time observations with no scooter trips. Separate comparisons of pilot, early-operation and later-operation periods tested whether relationships changed as each service matured.
Findings
The five cities shared broad demand rhythms and similar trip characteristics, yet the level and timing of use changed within cities as services developed. Weather, transport infrastructure, neighbourhood characteristics and public-transport accessibility were statistically associated with demand, and the strength of these relationships differed by city and service stage. The findings show why a launch-period trip count provides an incomplete benchmark: the same service can settle into different patterns as seasons, supply and local travel conditions change.