Transportation network company wait times in Greater Seattle, and relationship to socioeconomic indicators

Transportation network companies (TNC) which use smartphone apps to connect travelers to drivers for point to point, intra-city trips have expanded rapidly in recent years, yet the impacts of their services on urban transportation systems, for good or ill, are not fully understood. These services may increase access to transportation for some individuals, but it is unclear whether these benefits are equitably distributed across different neighborhoods. In this paper, we explore the spatial variability in wait times for a TNC vehicle throughout the Seattle, WA region, testing whether areas with lower average income or a greater percentage of minorities experience different waiting times than other areas. We collected approximately 1 million observations of estimated waiting times, quasi-randomly sampled over approximately two months in 2015. We analyzed spatial and temporal patterns using local regression, which suggested lower wait times in densely populated areas of the Seattle region, and the lowest wait times during midday hours. We developed a spatial error regression model to investigate relationships between wait times and socioeconomic indicators at the census block group (CBG) level. We find that conditional on other covariates, expected waiting times are longer in CBGs with higher average income, and shorter in CBGs with greater population density and employment density. After adjusting for differences in density and income, a higher percentage of minorities in a CBG is associated with longer waiting times late at night and shorter waiting times during the day, with an average effect of close to zero. Geographically weighted regression indicates that the strength, and in some cases the sign, of these relationships varies throughout the Seattle region. Overall, the results suggests that transportation network companies offer higher performance in dense urban areas, and that adequate access to TNC services is not necessarily restricted to areas that are “white and wealthy.”