On the determinants of Uber accessibility and its spatial distribution: Evidence from Uber in Philadelphia

摘要

Abstract This study investigates the impact of socioeconomics and demographic factors (e.g., population density, minority rate, age, gender, education, wealth, and crime) and transportation infrastructure (e.g., walk score, transit score, and bike score) on the accessibility of Uber in the city of Philadelphia. K ‐means clustering is applied for initial data exploration. Based on the spatial model selection diagnostic tests, we developed maximum likelihood spatial lag models with queen contiguity spatial weight matrix. The results show that Uber accessibility is not balanced in different neighborhoods in Philadelphia. Uber is more accessible in denser areas with the high male population, better public transportation access and less access to amenities in the walkable distances. Moreover, we observed that Uber is more accessible in areas with a high crime rate. This observation shows that Uber has made it easier to get out of high crime rate areas. Finally, contribution in the literature on accessibility in ride‐sourcing networks is discussed. Findings are additionally used to provide managerial implications to mitigate discrimination in ride‐sourcing platforms. This article is categorized under: Application Areas Industry Specific Applications Algorithmic Development Spatial and Temporal Data Mining Commercial, Legal, and Ethical Issues Social Considerations