Figure 1. In order to reveal hidden information and patterns in motion data (a), we propose two sequential techniques: the geospatial (b) and the temporal abstraction (c). The geospatial abstraction uses a density-based approach to cluster very dense areas based on user chosen properties. The temporal abstraction segments and visualizes the results of the geospatial abstraction as a graph network, represented in several small multiples. Through an interactive feedback loop the user is able to explore and analyze the trajectory.