A comparison of time-use behavior in metropolitan areas using pattern recognition techniques

Abstract

A better understanding of the connection of urban forms and travel behavior is critical to the operation of the existing and the design of the future transportation infrastructure. A comprehensive analysis of travel behavior across different regions would capture the underlying dependency among time-use behavior, the built environment, and the demographics of travelers. This paper introduces a method to measure the similarity of activity chains in different regions based on pattern segmentation and recognition. Travel behavior of residents of five different urban regions in the United States are compared: New York City, Los Angeles County, Chicago, San Francisco, and Atlanta. A total of 80,894 activity patterns is analyzed to address three goals: (1) to find a set of representative activity patterns for the residents of each region; (2) to analyze the dissimilarities/similarities in the activity patterns within and between the regions; and (3) to develop econometric models to assess and compare the role of demographic attributes on time use behavior. The outcome of the analysis supports and highlights the differences between study areas.