Visual Exploration of Large Multidimensional Trajectory Data

Abstract

Several visualisation methods have been recently proposed to aid a wide variety of users in the exploration of geographical trajectory, or trail, datasets. Such datasets consist of thousands up to millions of spatio-temporal trails that are also attributed by many additional data variables related to the identity of the tracked items, type of motion being recorded, data provenance, and more. As both data size and data dimensionality grow, finding efficient and effective ways to answer concrete questions, as well as discover unknown insights, from such data become increasingly important. We present an overview of recent information visualisation and visual analytics developments in this direction, with the aim of bridging the gap between Technical developments in this area and actual users and use-cases that can benefit from them. In this overview, we discuss strengths, limitations, assumptions, and other important characteristics of such visualisation methods, so as to help domain experts find optimal methods for their given application contexts. We illustrate our discussion with several examples of visualisation of large-scale, real-world, trajectory datasets related to migration data and use-cases.