Two-Stage Double Bootstrap Data Envelopment Analysis for Efficiency Evaluation of Shared-Bicycle Stations in Urban Cities

To optimize the performance and operation of shared-bicycle systems, this study aims to evaluate the efficiency of shared-bicycle stations and to find factors affecting their efficiency scores. We analyzed the efficiency of 1,260 shared-bicycle stations in the City of Seoul using shared-bicycle rental and trajectory data as of June 2018. In this study, the two-stage bootstrap data envelopment analysis, which is a non-parametric frontier technique, was applied to estimate each shared-bicycle station’s efficiency. In the first stage, we evaluated efficiency scores by employing the number of bicycle racks and bicycle path ratio as input variables, and bicycle turnover rate and balancing rate as output variables. The efficiency scores were regressed on potential covariates using a bootstrapped truncated regression in the second stage. From our results, the efficiencies of shared-bicycle stations were found to be diverse depending on the nature of land use around the station location. The results present evidence to show that shared-bicycle stations located in residential and school-dominated areas are likely to be efficient, whereas those in semi-industrial areas, commercial, and business districts are generally inefficient. Furthermore, the effect of variables like the commuting population, the number of registered vehicles, and the number of bicycle-related accidents per year were statistically significant, thus affecting shared-bicycle station performance. This study offers essential insights into the efficiency of shared-bicycle stations, which could be incorporated into shared-transportation strategies to improve mobility in urban cities.