Using Criteria Reconstruction for Low-sampling Trajectories as a Tool for Analytics

Abstract

Abstract Mobile applications equipped with Global Positioning Systems have generated a huge quantity of location data with sampling uncertainty that needs to be handled and analyzed. Those location data can be ordered in time to represent trajectories of moving objects. The data warehouse approach based on spatio-temporal data can help on the analysis. For this reason, we consider the problem of personalized reconstruction of low-sampling trajectories and include the criteria of movement as a dimension of analysis in a trajectory data warehouse. We enhance the analytics using dimensional modelling and graphical analysis in order to provide mechanisms to help decision makers. For example, analysts may formulate queries such as What are the top 5 most traversed streets between 07:00:00 am and 09:00:00 pm on August 9, 2014 (Saturday) if the trajectories are reconstructed using the touristic criterion? The answer to this query may help users to identify, e.g., city bottlenecks.