Optimising Representational Mix in Multi-Representational GeoVisualisation Contexts

Abstract

This thesis explores various ways to communicate the complexity of sentient movement visually. It argues that space, time and activity are three main dimensions of sentient movement, but activity is a privileged dimension for representing and visualising the movement of sentient objects. Sentient beings have goals and desires, which can be realised through their activities as they move through and interact with their environment and other sentient beings. The movement of sentient objects is the focus of this research due to its inherent importance for our understanding of this dynamic world, and the opportunities and challenges that have arisen from the increasing availability of large volumes of high resolution data for the movements of sentient animals and humans. To explore interactions between space, time and activity, this research examines various generalisation approaches that emphasise different combinations of space, time and activity dimensions as well as their properties. New generalisation techniques such as ringmaps and sta-rods are specifically developed to fully integrate space, time and activity in visualisations of the movement of sentient objects. The complexity and richness of sentient movement data demands coordinated multiple view visualisation techniques to draw together space, time and activity. This research presents a prototype visualisation toolset, called RINGMAPS, to visualise multiple patterns of sentient movement in an integrated way at different scales. The toolset has been applied to three data sets that are conceptually similar but contextually different in order to gain insights into sentient movement and to validate the visualisation techniques. In order to assess the visualisation toolset, a panel of four domain experts, each one familiar with one of the data sets, was invited to participate in an insight-based domain expert evaluation. The outcome of the evaluation of the three case studies suggests that the visualisation toolset is reasonably generic, and works well with sentient movements that are associated with different processes and different contexts. Coordinated multiple view visualisations that show integrated patterns of space, time and activity are effective and can provide deep insights. The degree of effectiveness is linked to factors such as scale, generalisation, the geographic context, and the nature of the data and their rhythms. Keywords: space time and activity, movement of sentient objects, time geography, generalisation, ringmap, coordinated multiple views, geovisualisation, insight-based domain expert evaluation.