Deriving Locations for a Station-Based Micromobility Sharing System from Historical Free-Floating Data: A Spatial Clustering Approach

This paper presents a methodology to derive potential locations for a station-based micromobility-sharing system. ~100,000 historical trip data of the MORE sharing system on the Campus of the University of the Bundeswehr Munich of a free-floating operating phase from begin of March 2023 to the end of February 2024 are subjected to spatial cluster analyses to identify potential locations. The trip data are clustered using four algorithms (k-means, k-medoids, fuzzy-c-means, and linkage). The best solution, which minimizes the number of stations while maintaining a defined level of spatial accessibility, is generated with the k-means algorithm. The potential locations must be evaluated for the actual station network design because different points like wifi coverage, vehicle parking facilities, and zone accessibility must be considered to define the actual sharing zones. The actual sharing zones minimize the operators’ costs to maintain the system while providing an optimal (or good enough) user experience.