Visual Statistics: Seeing Data with Dynamic Interactive Graphics

Iain Lang · Journal of the Royal Statistical Society Series A (Statistics in Society) · 2007

This book is based on the idea that graphics can be used to understand statistics, and to explore data, in ways that are within the grasp of those who lack statistical training. The idea is not new, and the field of graphical statistics is well developed, but the authors of Visual Statistics draw their inspiration from—alongside more typical sources—movies and video games, and emphasize the importance of graphics that are interactive and dynamic. By this they mean graphics that are fluid and responsive to the actions of the analyst, and that can be instantaneously manipulated on screen. An example of this would be a three-dimensional scatterplot that the analyst moves round in realtime; the authors suggest how data of various types may be played with and assessed in this way. They also discuss ways of using the approach to deal with missing data and in the visual fitting of regression models. A central component of the book is a piece of software called ViSta: the Visual Statistics System, which is freely available on line. Most of the concepts that are covered in the book are set out and explained in a general way, but the examples that are presented refer to ViSta and this makes the software key to the usefulness of the book. I found the book's presentation clear and straightforward, but the software package and the Web site from which one can download it are less well designed and maintained. It is not clear which version of the software is the latest (versions 5.5 and 6.4 are both referred to), the software is said to have been tested on Windows 95, 98, NT4 and 2000 (without mention of the newer versions) and there are dead links within the site. The software itself seems powerful and sufficiently flexible to be useful for teaching purposes, but suffers, like many non-commercial packages, from the fact that it is difficult to import data sets from other mainstream packages. It can handle files in Microsoft EXCEL format but users must otherwise enter data by hand or import them as text. On the positive side the site's discussion group seems to be active, which bodes well for future development. Those who want to know whether the book will be suitable for their needs are recommended to take a look at the Web site, and to download the software, for themselves.

Read the paper · More papers on PaperTik