Small Steps, Big Differences: Assessing the Validity of using Home and Work Locations to Estimate Walking Distances to Transit

Document Type

Journal Article

Publication Date


Subject Area

mode - pedestrian, place - north america, place - urban, planning - surveys, ridership - behaviour, ridership - mode choice, technology - passenger information


physical activity, public transport, walking distance to transit


Walking to and from public transport can form a seamless way to integrate physical activity into our daily lives, thereby helping us achieve the recommended minutes of physical activity. To measure the link between physical activity and public transport use, it is critical to determine how far individuals are walking, and are willing to walk, to different modes of transit. Few planners, however, have access to detailed information on the exact public transport lines used by individuals, and therefore need to estimate distances by making use of only home and work locations. This study therefore compared two methods of calculating walking distances: one method using widely available home and work locations and a fastest route algorithm leveraging general transit feed specification data, and a second employing a detailed travel survey containing information on the real routes used by each respondent from Montreal, Canada, to generate more robust estimates of the distances individuals are walking to public transport stops. Results show that walking distances calculated from commonly available origin and destination information tend to underestimate real walking distances by 10%. Multilevel mixed-effect regression models indicate these differing results are mainly attributable to differences in travel behaviour and mode choice. Findings from this study provide a better understanding of how modelled and real walking routes to public transport stops differ, which could be of interest to professionals and urban decision makers wishing to correctly model walking to transit in their region when only limited information is available.


Permission to publish the abstract has been given by SAGE, copyright remains with them.