How Equitable is Connected Vehicle Data for Segment-based Performance Measures?

Abstract: In the past few years, commercial vendors have begun to sell connected vehicle (CV) data aggregated from a variety of sources such as auto manufacturers. With such data necessarily originating from equipped vehicles, there is a potential that such data may have better coverage in locations where newer vehicles are likely to be owned by residents or employees. This study explores the influence of such variables on the penetration rate for this type of data. The study employs Geographic Information System (GIS) data and spatial analysis methods to explore the influence of local median household income and other variables obtained from the American Community Survey. The count of connected vehicle data was also obtained using GIS analysis, and a penetration rate was calculated by comparison against AADT for the same segments. Regression models were used to establish a relationship between the CV penetration rate and the socioeconomic variables. The results show that the penetration rate was higher in high-income areas than in low-income areas. In addition, locations with larger disadvantaged populations had lower penetration rates. These effects were found to be statistically significant. However, a marginal effects analysis found that the magnitude of the effects were small. The low magnitude of the effects are encouraging for the possibility of developing equitable segment-based performance measures, yet the significance of the effects serve as a reminder that it is important to account for the possibility of data underrepresentation