Transformations, Mappings, and Data Summaries

Abstract

Fundamentally, data visualization is the process of placing dabs of ink or color on a 2D plane. However, the complexity of data is increasing so that we see large numbers of instances, dimensions, parameters, etc. Such data surpasses what can readily be shown on a 2D or 3D display. One solution to this challenge is the development of better or more complex interfaces, that include, for instance, linked views, large displays, dynamic visualizations, and sophisticated user interactions. The alternative and complementary approach is to develop sets of mathematical and statistical tools to transform, map, or summarize data and thereby reduce its complexity so that visualization and understanding of large and complex becomes more feasible. The role of visualization research, in this case, is to identify common use cases and develop methods and tools that can readily be adapted to particular applications. To address the challenges of complexity in the data, previous works have proposed reducing items and attributes and associated visualization conventions and practices. Here we take deeper (and complementary) look at the analytical frameworks and approaches for transforming data into forms that are appropriate for display devices, considered generally. The approach in this chapter is to begin by characterizing different types of data in a way that is well suited for this discussion. We will then focus on a few particular classes of data and different ways of summarizing and transforming data of those types. Finally, we will broaden the discussion to other types of data and how they map into the various methodologies.