A model-based approach to select case sites for walkability audits

Walkability audits provide valuable information about pedestrian environments, but are time-consuming and can be expensive to implement. In this paper, we propose a model-based approach to select sites for conducting walkability audits. The key idea is to estimate a model of travel behavior at the meso-scale level, which can be examined to identify locations where the behavior is under- and over-estimated. We conjecture that systematic under- and over-estimation can be caused by micro-level factors that influence the behavior. The results can be used to identify sites for walkability audits. The approach is demonstrated with a case study in Hamilton, Canada. A model of walk shares forms the basis of the site selection procedure. After identifying areas with higher and lower shares than predicted by the model we select a sample of neighborhoods for audits. Analysis of the results reveals elements of the local environment that associate with greater-than-expected walk shares. The case study demonstrates that the proposed model-based strategy can be used to better target limited resources, and produce valuable insights into micro-level factors that affect travel behavior.