Abstract: Commuting is one of the most important parts of travel behavior, which has been widely studied by geographers, planners and transportation researchers based on questionnaire data, and the rapid development of location-based technologies and ICTs bring new opportunities for deep insight of commuting with tracking data. The study focuses on variability in commuting behavior in different weekdays, introduces the concept of activity flexibility and brings forward a concept of commuting flexibility. Four dimensions, i.e., time, space, travel mode and route choice, are used to define the concept of commuting flexibility, and seven theoretical commuting patterns are put forward according to discussion on the relationship between four dimensions of commuting flexibility. The study takes two suburban mega-communities in Beijing, i.e., Tiantongyuan and Yizhuang, as study cases. It collects seven days of space-time data of individual behaviors through a web-based survey, which is a combination of activity-travel dairies and GPS tacking data. On the basis of the analysis of commuting behavior, the study explores measures and characteristics of four types of commuting flexibility and represents the activity-travel trajectories in a 3-D GIS environment. The concept of commuting flexibility proves to be effective and reasonable for a better understanding of complicated commuting behaviors of Beijing's residents. The study also provides a special perspective for Beijing's urban and transportation problems.

Key words: GPS, commuting flexibility, commuting pattern, 3-D visualization, Beijing