Modeling Walk Trips Using a Generalized Accessibility Framework

This paper develops and applies a generalized framework for extending accessibility measurements from the metropolitan to the walking scale, directly addressing a key limitation in much of the current research and practice in transportation planning and modeling. It integrates spatially distributed activities into a multi-modal transportation graph, allowing the efficient analysis of destinations at multiple scales by transportation mode. This research makes several methodological advances that enable representation of the full set of local streets for pedestrian scale accessibility, a hierarchical graph to represent the tradeoff between modes, and integration of micro-scale land use data to measure the full population of alternative destinations in the city. This paper focuses on pedestrian trips using the 2012 California Household Travel Survey (CHTS) as an application of this framework, and compares the predictive power of Walkscore to other accessibility measures found in the literature, as well as household and individual attributes The results demonstrate the utility of using a framework for computing accessibility at mode-appropriate scales for predicting travel outcomes such as walking trip counts. The authors find that accessibility variables are correlated, and that composite metrics such as Walkscore provide a parsimonious predictor of walking trips, mediated by household and individual attributes. The results confirm that local residential density variables and regional accessibility to jobs, help explain walking trips, even after controlling for Walkscore. In addition, regional accessibility by other modes appears to complement local accessibility in increasing walk trips.