Abstract
Urban commercial districts and centers are places that provide concentrated opportunities for non-work activities. Rapid development in these areas has made them critical for local economic development as well as exerting significant influence on urban society and culture. Traveling to these non-work destinations, such as shopping centers, restaurants, bars, grocery stores, movie theaters, etc., is an important part of urban life. For a long time, survey-based data is often used to examine non-work trips and travel patterns. These data always have limited sample sizes that impede temporally and spatially fine-grained analysis. Recent advances in information and communication technology (ICT) and mobile devices create new opportunities for today’s transportation planners to understand travel behavior using non-survey sources of data. These data are user-generated, geo-located, and contain contextual information (e.g., text, images, videos). The emergence of such “transportation big data” has resulted in a large quantity of information documenting people’s everyday movements, travel events, attitudes, perceptions, and emotions, all connected with the location and time.
This dissertation develops a data fusion framework that integrates geosocial media, fine-grained individual GPS trace data, land use and built environment data, and demographic data from the U.S. census to quantify people’s travel experiences and mobility patterns to commercial and mixed-use districts, taking the Phoenix Metropolitan Area as a study case. Specifically, the geosocial media data used in this dissertation is collected from Yelp reviews and the GPS trajectory data is collected from smartphone apps with GPS-enabled location services. This dissertation research first examines the experience of travel (travel attitude) in major commercial and mixed-use districts using transportation texts embedded in Yelp reviews. Then, it analyzes travel behavior to these destinations using GPS trajectory data with a fine scale in space and time. Following on from the prior two analyses, it develops a data fusion framework by integrating geosocial media and GPS traces to further examine 1) the relationship between attitude and built environment, and 2) the impacts of attitude and built environment on travel behavior.
Given the prospect of the big data era for transportation research, this dissertation research shows the promises of emerging data and analytics in providing useful information about travelers’ attitudes and behaviors. It also enhances our understanding of non-work travel and has implications for transportation planning and management. Therefore, this dissertation makes two major contributions to urban transportation planning research, one regarding the travel to non-work destinations, and second regarding the methods developed to integrate multiple types of big data for transportation planning informatics.
