Exploratory visual pattern detection of mobile object data in attribute-time space

Recent years witnessed the emergence of massive individual-based movement data due to the location-aware devices such as global positioning system (GPS), mobile phones and radio-frequency identification (RFID). These data are overwhelming the techniques of traditional spatial analytic techniques. Researchers and practitioners are turning to spatio-temporal knowledge discovery and exploratory visualization techniques to find patterns, trends and relationships hidden in the large volume mobile object datasets. This research develops a method to visualize mobile object data in both space-time and attribute-time to discover hidden knowledge about the evolution of dynamic path properties in concert with its location in space with respect to time. Respatialization techniques project mobile object trajectories from geographic space and time to a multivariate space and time defined by choosing three other dynamic attributes of the path and forming a space from the cross-product of these attributes with time. The attributes may be other spatial or geometric properties of the path, or non-spatial quantitative attributes that are dynamic. Dual visualizations of trajectories within space-time and attribute-time can provide new insights to the dynamic evolution of individual and collective spatio-temporal patterns. In addition, visual summarization of attribute-time trajectories by three-dimensional convex hulls provides intuitive and quantitative comparison of trajectories. A visualization software environment that implements these concepts with a case study application to empirical trajectory data presents the effectiveness of the exploratory visual pattern detection in mobile object data.