TrajAnalyst: Matching Data to Trajectory Analysis Modules via a Conceptual Framework

With the fast growth of location awareness devices and sensors (e.g., GPS, mobile phones and RFID tags) as well as location-based social networks (e.g., Foursquare, Jiepang), large-scale movement data (or trajectories) have become available to GIScience community and related domains. Researchers have developed amount of theories, models and techniques for movement pattern analysis (MPA) and applied in many interdisciplinary/multidisciplinary studies, including but not limited to investigating human mobility (González et al. 2008, Song et al. 2010, Kang et al. 2012), travel behaviour and traffic flow estimation (Gao et al. 2013, Liu et al. 2013), animal interactions (Miller 2012), and hurricane trajectory similarity (Dodge et al. 2012). The movement trajectories can be generalized as a series of spatiotemporal points with diverse contexts. As trajectory data often share common structures and semantics, such as location fix, order, timestamp, segment, direction, speed and so on (Hu et al. 2013), a number of generic analysis methods reoccur in the trajectory data processing across different domains. However, these generic analysis modules and methods are still isolated from GIS integration perspective, although there is a large group of researchers using R package (Kranstauber and Safi 2014) or ArcMET 1 for movement patterns in animal ecology. Previous trajectory studies primarily focused on space-time integration and visualization but little on generic functionality. We still lack tools supporting trajectory analyses (Goodchild 2013). Thus, the motivation of this research is to take a first-step towards creating an integrated trajectory analysis framework and tools which can be plugged in a widely used GIS platform to better support the processing and spatial analysis of trajectory data. A theoretical question is how to match trajectory data to corresponding functionality. We present a conceptual framework for linking the data structure to the analyses and supporting better human-computer interaction.