Applying data science to analyze internal trip capture for station area neighborhoods

• We examine internal trip capture with big data in Portland and DC neighborhoods. • We find generally high internal trip capture using big data, 41-80%. • Employment density is most strongly related to internal trip capture. • Regional transit accessibility is also related to internal trip capture. • Various mixed-use variables have minimal additional correlation. Local accessibility, or accessibility by proximity, is a planning priority that is increasingly discussed within the context of the 15-Minute City. The goal of local accessibility is both to facilitate residents’ quality of life and to promote more sustainable travel, especially through nonmotorized modes. Here, we examine the degree of local accessibility provided by transit neighborhoods by examining internal trip capture. Much of the previous work on internal trip capture has focused on mixed-use developments, yet the same methodology has been adopted at larger scales and is relevant to the 15-Minute City concept. We examine 1.6-km (1-mile) neighborhoods around 96 transit stations in Portland, Oregon and 96 transit stations in metropolitan Washington, DC, both in the US. The outcome variable of interest is the percentage of internal trip capture, defined as the percentage of trips with both trip origins and trip ends located within these neighborhoods. Our data on trips comes from the big data vendor, StreetLight; therefore, our analysis is based on thousands of trips per study area rather than dozens, in contrast to previous internal trip capture studies. The independent variables encompass a wide range of built and social environment variables, including various measures of street network connectivity, the concentration of jobs and population, jobs disaggregated by industry, a variety of mixed-use variables, median income, and accessibility to transit and auto jobs. We find that job density, street network connectivity, and transit job accessibility all correlate with higher levels of internal trip capture in both settings, with internal trip capture rates ranging from 41% to 80%.