A Methodology on Quantifying Spatial-Temporal Disparities in Travel Behavior of Mobility-Impaired Travelers with Non-Disabled Travelers: Case Studies of London and New York

Abstract

Background Physically-impaired citizens are, arguably, the most vulnerable group of the society and, unfortunately, often the least favored one. Transportation and mobility are key players in determining the qualities of life such that transport disadvantaged urban forms are associated with the risk of social exclusions of the citizens. For physically-impaired individuals, the burden caused by an inaccessible transportation system is more severe due to the coupled effects of limited mobility due to impaired personal health and lack of good mobility service. Providing a good and accessible transportation service requires (1) a better understanding of travel behavior of the target population, (2) quantifying mobility disparities between the patterns of the target population and others, and (3) detecting the operational gap. Methods We developed a methodology to compare the mobility behavior in two dimension of time and space. The steps are: Generation of Scheme Characteristic Vectors (SCV) to compare the patterns in the dimension of space. These vectors are generated based on some set of geometric and shape formatting operations; Generation of Temporal Activity Strings (TAS) to compare the patterns in the dimension of time; Clustering SCV and TAS in the dimensions of time and space using a two stage clustering method: affinity propagation and k-means clustering; Finding the K-nearest neighbors (KNN) based on the demographic attributes to measure the dissimilarity in the observed mobility behavior of physically disabled travelers to others who have similar demographic attributes. This step addressed the question of ‘what would be the spatial-temporal pattern of a disabled person if she/he was not disabled?’ Results We applied the methodology to London and NYC. A total of seven temporal clusters for NYC and 10 temporal clusters for London emerged from the data. For spatial analysis, 10 spatial patterns for NYC and 13 patterns for London were formed. In both regions, a comparison of the time-use behavior and spatial mobility shows significant disparities in time-use patterns of disabled people. A lower percentage of disabled travelers spend their time out of home, or make short trips. In spatial pattern comparison, the difference between mobility patterns in NYC is more significant than in London. Conclusions We developed innovative techniques to quantify the gap between the mobility behavior of disabled people and others. The results can be used in any urban planning decision making, where multiple factors of economic considerations and equity need to be taken into account.