Spatial disparities of Uber accessibility: An exploratory analysis in Atlanta, USA

- Spatial disparities of accessibility in Atlanta, USA are explored using spatial autoregressive models with Uber taxi data.
- Accessibility is measured as the expectation and variability of Uber wait times using UberX and UberBLACK.
- Wealth and race do not have significant associations with Uber accessibility.
- Higherroad network density, population density and less commuting times correlate with greater Uber accessibility.

Inequality of accessibility in transportation systems is a constant concern, which is intensified by the transportation economization process and the digital divide. How should the accessibility of crowdsourced transportation is measured and understood? Without any prior assumption, this paper openly explores spatial disparities of accessibility in the city of Atlanta, USA using both the UberX (the most popular Uber product) and the UberBLACK (the premium Uber product) data. Accessibility is measured by both the expectation and variability of Uber wait time. With spatial autoregressive models, we find that after controlling for other socioeconomic factors, wealth and race do not have significant associations with Uber accessibility. Additionally, higher road network density, population density, and less commuting time to work correlate with greater Uber accessibility. More public transport stops are related to better accessibility of UberX but worse accessibility of UberBLACK. Finally, implications for policy-makers are provided.