Interactive and dynamic statistical graphics enable data analysts in all fields to carry out visual investigations leading to insights into relationships in complex data. Interactive and dynamic statistical graphics involve methods for viewing data in the form of point clouds or modeled surfaces. Higher-dimensional data can be projected into one-, two- or three-dimensional planes in a set of multiple views or as a continuous sequence of views which constitutes motion through the higher-dimensional space containing the data.
Strictly, there is some difference between interactive graphics and dynamic graphics. When speaking of interactive graphics only, we usually mean that a user actively interacts with, i.e., manipulates, the visible graphics by input devices such as keyboard, mouse, or others and makes changes based on the visible result. When speaking of dynamic graphics only, we usually mean that the visible graphics change on the computer screen without further user interaction. An example for interactive graphics might be the selection of interval lengths and starting points when trying to construct an optimal histogram while looking at previous histograms. An example for dynamic graphics might be an indefinitely long grand tour with no user interaction. Typically, interactive graphics and dynamic graphics are closely related and we will not make any further distinction among the two here and just speak of interactive and dynamic statistical graphics.
Two terms closely related to interactive and dynamic statistical graphics are exploratory data analysis (EDA) and visual data mining (VDM).
EDA, as defined by [167], ''is detective work - numerical detective work - or counting detective work - or graphical detective work.'' Modern techniques and software in EDA, based on interactive and dynamic statistical graphics, are a continuation of Tukey's idea to use graphics to find structure, general concepts, unexpected behavior, etc. in data sets by looking at the data. To cite [167] again, ''today, exploratory and confirmatory can - and should - proceed side by side.'' Interactive and dynamic statistical graphics should not replace common analytic and inferential statistical methods - they should rather extend these classical methods of data analysis.
Data mining (DM) itself ([198,109]), see also Chap. III.13, is a field whose scientific basis has only began to be explored over the last few years. DM exists as a result of the convergence of several fields including data bases, statistics, and artificial intelligence. [77] discusses the connection between DM and statistics in more details and [183] provides a definition of DM that links it with EDA and graphics: ''Data mining is exploratory data analysis with little or no human interaction using computationally feasible techniques, i.e., the attempt to find interesting structure unknown a priori.'' Simultaneously with an increasing interest in DM there has been the evolution of computer graphics, especially in the area of virtual reality (VR). Within the statistics framework, the area of EDA has evolved into a more sophisticated area of interactive and dynamic statistical graphics. Recently, DM has been combined with statistical graphics, resulting in VDM ([59,103,149,114,16]). However, there exist several different definitions of the term VDM. [139] dedicate less than one page to ''dynamic visualizations that allow user interaction'' in their book on VDM.
In this chapter we will provide a general overview on existing methods and software for interactive and dynamic graphics. We will also provide a snapshot of current developments that may become a standard in the near future but may also be quickly forgotten again. All sections are supported by an extensive list of references that will allow every reader from novice to expert to become more familiar with a particular concept of interactive and dynamic graphics. More specifically, in Sect. 10.2, we will discuss early developments and software related to interactive and dynamic graphics. In Sect. 10.3, we will discuss the main concepts and in Sect. 10.4 some software products related to interactive and dynamic graphics. Interactive 3D graphics will be discussed in Sect. 10.5 and applications of interactive and dynamic graphics in geography, medicine, and environmental sciences will be discussed in Sect. 10.6. We conclude this chapter with an outlook to possible future developments in Sect. 10.7.
All graphical displays throughout this chapter are based on the ''Places'' data set that was distributed to interested members of the American Statistical Association (ASA) several years ago so that they could apply contemporary data analytic methods to describe these data and then present results in a poster session at the ASA annual conference. The data are taken from the Places Rated Almanac ([18]). The data are reproduced on disk by kind permission of the publisher, and with the request that the copyright notice of Rand McNally, and the names of the authors appear in any paper or presentation using these data. The data consist of nine measures of livability for 329 cities in the U.S.: Climate and Terrain, Housing Cost, Health Care and Environment, Crime, Transportation, Education, The Arts, Recreation, and Economics. For all but two of the above criteria, the higher the score, the better. For Housing Cost and Crime, the lower the score the better. The scores are computed using several statistics for each criterion (see the Places Rated Almanac for details). Latitude and longitude have been added by Paul Tukey. Population numbers have been added as well.