Editing, Writing, and Translation

Home Services Books Articles Resources Fiction Contact me Français

You are here: Articles --> 2013 --> Book review: Data representations, transformations, and statistics for visual reasoning
Vous êtes ici : Essais --> 2013 --> Book review: Data representations, transformations, and statistics for visual reasoning

Data Representations, Transformations, and Statistics for Visual Reasoning

Ross Maciejewski. 2011. San Francisco, CA: Morgan & Claypool Publishers. [ISBN 978-1-60845-625-3. 78 pages. US$35.00 [softcover].)

Previously published as: Hart, G. 2013. Book review: Data representations, transformations, and statistics for visual reasoning. Technical Communication 60(2):171.

Most STC members have heard of Edward Tufte’s classic book, The Visual Display of Quantitative Information. His work is a great starting point for exploring data graphics whether you’re a scientist or an educated amateur. In Data Representations, Ross Maciejewski looks at data graphics from the perspective of those who actually create and manipulate the data, including scientists, engineers, and statisticians. As a result, this book is not for the faint of heart. (Warning: Here be equations!)

Where Tufte relies on persuasive visual examples to make his case, Maciejewski digs deep into the mathematical characteristics of data. In this small book, he explains the different data types (e.g., nominal versus ordinal), the various uses of colors to represent value ranges, what he calls “preconditioning” (transformation of the data so that its distribution meets the needs of statistical analysis), how to select appropriate “bins” (value ranges), and how these factors affect the main graph types you can use to display data more effectively. The goal is to help data mavens understand their own data, present it to others, and use it to support reasoning about the meaning of the data. Perhaps the best part of Data Representations, though it’s left almost entirely implicit, is the recognition that thinking about and creating data graphics are still largely bound by the constraints of the print model (and its PDF descendant). Maciejewski reminds us that by designing visualization tools that can be shared over the Web, we move beyond sharing our data with our reader to encouraging our reader to interact with that data. That’s a paradigm-changing insight that deserved more explicit treatment.

Unfortunately, the writing is often dense and includes occasional errors (“with green being the lease severe alert”, p. 8; unlabeled graph axes, p. 14) and many terms that assume prior knowledge and that would benefit from explanation (“spread variation”, p. 12), and this is exacerbated by the use of fussy, too-small type that makes the book unnecessarily difficult to read. There are also poor choices such as the failure to use any colors other than grayscale in the chapter on color use—not even a link to a page of color images on the publisher’s Web site, as is commonly done for research journals that don’t publish color printed versions. (The bibliography is decently long, but no Web links appear there or in the text.) The book clearly could have used a more rigorous developmental edit.

These factors may scare off some potential readers of Data Representations, although most of the intended audience will be able to plow through and understand the many important points. If you’re not a member of this audience, but willing to devote the necessary effort, you’ll develop an understanding that provides considerable credibility when you discuss data graphics with the subject-matter experts who create them.

©2004–2017 Geoffrey Hart. All rights reserved