This book is based upon the extensive practical experience of the authors in designing and developing software tools for visualisation of spatially referenced data and applying them in various problem domains. These tools include methods for cartographic visualisation; non-spatial graphs; devices for querying, search, and classification; and computer-enhanced visual techniques. A common feature of all the tools is their high user interactivity, which is essential for exploratory data analysis. The tools can be used conveniently in various combinations; their cooperative functioning is enabled by manifold coordination mechanisms.
Typically, our ideas for new tools or extensions of existing ones have arisen from contemplating particular datasets from various domains. Understanding the properties of the data and the relationships between the components of the data triggered a vision of the appropriate ways of visualising and exploring the data. This resulted in many original techniques, which were, however, designed and implemented so as to be applicable not only to the particular dataset that had incited their development but also to other datasets with similar characteristics. For this purpose, we strove to think about the given data in terms of the generic characteristics of some broad class that the data belonged to rather than stick to their specifics.
From many practical cases of moving from data to visualisation, we gained a certain understanding of what characteristics of data are relevant for choosing proper visualisation techniques. We learned also that an essential stage on the way from data to the selection or design of proper exploratory tools is to envision the questions an analyst might seek to answer in exploring this kind of data, or, in other words, the data analysis tasks. Knowing the questions (or, rather, types of questions), one may look at familiar techniques from the perspective of whether they could help one to find answers to those questions. It may happen in some cases that there is a subset of existing tools that covers all potential question types. It may also happen that for some tasks there are no appropriate tools. In that case, the nature of the tasks gives a clue as to what kind of tool would be helpful. This is an important initial step in designing a new tool.
Having passed along the way from data through tasks to tools many times, we found it appropriate to share the knowledge that we gained from this process with other people. We would like to describe what components may exist in spatially referenced data, how these components may relate to each other, and what effect various properties of these components and relationships between them may have on tool selection. We would also like to show how to translate the characteristics of data and structures into potential analysis tasks, and enumerate the widely accepted principles and our own heuristics that usually help us in proceeding from the tasks to the appropriate approaches to accomplishing them, and to the tools that could support this. In other words, we propose a methodological framework for the design, selection, and application of visualisation techniques and tools for exploratory analysis of spatially referenced data. Particular attention is paid to spatio-temporal data, i.e. data having both spatial and temporal components.
We expect this book to be useful to several groups of readers. People practising analysis of spatially referenced data should be interested in becoming familiar with the proposed illustrated catalogue of the state-of-theart exploratory tools. The framework for selecting appropriate analysis tools might also be useful to them. Students (undergraduate and postgraduate) in various geography-related disciplines could gain valuable information about the possible types of spatial data, their components, and the relationships between them, as well as the impact of the characteristics of the data on the selection of appropriate visualisation methods. Students could also learn about various methods of data exploration using visual, highly interactive tools, and acknowledge the value of a conscious, systematic approach to exploratory data analysis. The book may be interesting to researchers in computer cartography, especially those imbued with the ideas of cartographic visualisation, in particular, the ideas widely disseminated by the special Commission on Visualisation of the International Cartographic Association. Our tools are in full accord with these ideas, and our data- and task-analytic approach to tool design offers a way of putting these ideas into practice. It can also be expected that the book will be interesting to researchers and practitioners dealing with any kind of visualisation, not necessarily the visualisation of spatial data. Many of the ideas and approaches presented are not restricted to only spatially referenced data, but have a more general applicability.
The topic of the book is much more general than the consideration of any particular software: we investigate the relations between the characteristics of data, exploratory tasks (questions), and data exploration techniques. We do this first on a theoretical level and then using practical examples. In the examples, we may use particular implementations of the techniques, either our own implementations or freely available demonstrators. However, the main purpose is not to instruct readers in how to use this or that particular tool but to allow them to better understand the ideas of exploratory data analysis.
The book is intended for a broad reader community and does not require a solid background in mathematics, statistics, geography, or informatics, but only a general familiarity with these subjects. However, we hope that the book will be interesting and useful also to those who do have a solid background in any or all of these disciplines.