Felice Frankel relishes an inspired handbook on the art and science of picturing data.
Data Points: Visualization That Means Something
- Nathan Yau
Scientists, probably more than most, are aware of the ever-increasing presence of data visualization in newspapers, television, online shopping, social media and even policy-making sales pitches in the US Congress. I'll bet there's a brand new office building in Washington DC devoted to creating chart-laden poster boards for congressional members. Statistician and visualization expert Nathan Yau's Data Points: Visualization That Means Something is a clear and passionate exploration of this burgeoning phenomenon.
A detailed handbook, Data Points is especially useful for those working on scientific data visualization, guiding the reader through fascinating examples of data, graphics, context, presentation and analytics. But this is more than a mere how-to manual. Yau reminds us that the real purpose of most visualization work is to communicate data to pragmatic ends. He points to the other end of the spectrum: visualizations created by those so seduced by artful design that meaning becomes inconsequential. As Yau writes of statistics and design knowledge, “having skills in both provides you with the luxury — which is growing into a necessity — to jump back and forth between data exploration and storytelling”.
Yau entices us to look and think, study, dissect and judge. As a visualizer himself, he has made intelligent choices of illustrations in this wonderfully varied collection; they are not there as decoration. Lokesh Dhakar's Coffee Drinks Illustrated (http://go.nature.com/tcxs21), for instance, is accessibly simple, yet rich in information. Yau includes an interactive astronomical visualization by Santiago Ortiz (http://go.nature.com/hwnsdx) as an example of how perspective and context are inseparable when zooming and rotating the night sky. The visualization of constellations is both elegant and beautiful, and could become an inspiration for scientists creating interactive molecular configurations. And although there are few specifically scientific visualizations on offer, the creative researcher could find much to adapt in these approaches.
Yau illustrates and discusses the fundamental components of visualizations and how small changes can improve readability. He teaches us to think graphically. On one spread (and there are many), he cleverly uses the 'visual cues' described by William Cleveland and Robert McGill in their 1985 study on graphical perception and methods (W. S. Cleveland and R. McGill Science 229, 828–833; 1985) in a table that subtly encourages the reader to think their way into visual abstraction (section pictured). The table shows patterns in data — such as an increase or decrease in the population of a species — and how these can be represented graphically through angle, position, area or colour saturation. How to visually abstract a concept or data into a formatted representation — that is, to reduce a visual expression to its fundamental information — is one of the least discussed but most important elements of successful visual representations.
I have some quibbles. You might need a magnifying lens while reading Data Points: a number of the figures need considerable enlargement to make sense on the printed page. And ensure you view online his many examples of visualizations intended for online publication, especially those intended for interactivity. I would also have liked to see a more in-depth discussion on representing uncertainty. We are all aware of the various levels of imprecision in our data, and not to communicate that somehow in our representations can be irresponsible — but that exercise probably warrants a separate book. Finally, I found it annoying that many figures do not appear next to where they are referenced in the text; I sometimes had to turn a page or even two to see what Yau was describing.
But this remains a masterwork. I can imagine some initial eye-rolling in the visualization community on the first viewing of some of Yau's seemingly 'obvious' examples. But there is much to learn from studying what Yau does here. That is, defining and demonstrating good visualization with clarity and precision as “a representation of data that helps you see what you otherwise would have been blind to if you looked only at the naked source ... trends, patterns, and outliers that tell you about yourself and what surrounds you”.
Thank you, Nathan Yau, for helping us to begin.
About this article
European Journal of Operational Research (2019)
Fit, Bias, and Enacted Sensemaking in Data Visualization: Frameworks for Continuous Development in Operations and Supply Chain Management Analytics
Journal of Business Logistics (2016)