Space-time classification of public transit smart card users' activity locations from smart card data

Smart card data from public transit systems has proven to be useful to understand the behaviors of public transit users. Relevant research has been done concerning: (1) the utilization of smart card data, (2) data mining techniques and (3) the utilization of data mining in smart card data. In prior research, the classification of user behavior has been based on trips when temporal and spatial classifications are considered to be separate processes. Therefore, it is of interest to develop a method based on users' daily behaviors that takes into account both spatial and temporal behaviors at the same time. In this work, a methodology is developed to classify smart card users' behaviors based on dynamic time warping (DTW), hierarchical clustering and a sampling method. A three-dimensional space-time prism plot demonstrates the efficiency of the algorithm.