An end to end framework for building data cubes over trajectory data streams

In this paper we propose an end to end framework that allows efficient analysis for trajectory streams. In particular, our approach consists of several steps. First, we perform a partitioning strategy for incoming streams of trajectories in order to reduce the trajectory size and represent trajectories using a suitable data structure. After the encoding step we build specialized cuboids for trajectories in order to make the querying step quite effective. This problem revealed really challenging as we deal with data (trajectories) for which the order of elements is relevant thus making the analysis quite harder than for classical transactional data. We performed several tests on real world datasets that confirmed the efficiency and effectiveness of the proposed techniques.