Visualising Multivariate, Spatiotemporal Data Effectively

Abstract

The amount of data with a temporal and geospatial component has increased rapidly over the past years under the influence of social media platforms like Facebook and Twitter. Visualisation of this data, especially with large datasets, can offer insight and the ability to find patterns and trends within the data. Glyph placement is an existing and effective technique for visualising multivariate geospatial data. In this study, we provide an overview of the literature on interactive data exploration and discuss existing approaches for placing glyphs in a 2D map. We formalise the problem of creating qualitatively good glyph maps as the Basic Glyph Placement (BGP) problem. We prove that this problem is NP-hard. We present a sound (though not complete) exact algorithm to solve it, inspired by Drezner's method for solving the continuous k-center problem. Additionally, we propose two approximation algorithms and evaluate their performance on three datasets. Finally, we present a mock-up of an interactive tool designed for visualising multivariate, spatiotemporal data.