Experience the city : analysis of space-time behavior and spatial learning

Learning plays an important role by coding information into individual cognitive maps that can be used to make decisions concerning individual behavior in space. Through traveling people learn about the urban environment and update their knowledge. In this regard, the growing concern in the field of urban planning and transportation modeling is to include information on traveler’s states of knowledge – their cognitive awareness or cognitive maps of the travel environment – for predicting more realistic travel behavior. However, it is not an easy process to describe and model behavioral changes and learning processes over time. The first reason is that the spatial learning process is a complex phenomenon that comprises different dimensions. The second reason concerns the availability of longitudinal empirical data. More specifically, two sources of data are required: (1) data on cognitive mapping process and (2) data on actual travel behavior. In the line with these considerations, this PhD thesis aims to study learning dynamics of newcomers in an urban environment based on the longitudinal activity-travel data collected over several weeks time horizon and complemented with data on cognitive process. For understanding the relationship between spatial knowledge acquisition and space-time behaviors of individuals a combination of qualitative and quantitative techniques for analyzing space-time behavior of a sample of newcomers to Eindhoven, the Netherlands has been used. Because the new generation of dynamic activity-based models requires multi-day or multi-week travel data, GPS tracking technology was utilized to collect data during several weeks on activity-travel patterns of newcomers. Much effort was put into developing and improving this technology for collecting data. These data were complemented by a series of consecutive interviews, internet surveys and maps drawing techniques. The major methodological contribution of this thesis is the development of an approach to imputing activity-travel patterns from GPS traces. Personal judgment and ad hoc rules are replaced by a Bayesian belief network, which represents the multiple relationships between spatial, temporal and other factors and the travel behavior aspects of interest, including errors in the technology itself. The results of analysis of activity-travelled patterns derived from GPS traces suggest that individuals possess different learning styles and activity space profiles. The findings allow to draw conclusions that information embedded in cognitive maps is strongly related to individual interactions within the city