Dave Choi writes:

I’m building a course called “Exploring and visualizing data,” for Heinz college in Carnegie Mellon (public policy and information systems). Do you know any books that might be good for such a course? I’m hoping to get non-statisticians to appreciate the statistician’s point of view on this subject.

I immediately thought of Bill Cleveland’s 1985 classic, The Elements of Graphing Data, but I wasn’t sure of what comes next. There are a lot of books on how to make graphics in R, but I’m not quite sure that’s the point. And I’m loath to recommend Tufte since it would be kinda scary if a student were to take all of his ideas too seriously.

Any suggestions?

I’d recommend Stephen Few’s “Show Me the Numbers” and “Now You See It”. Also, Naomi Robbins’ “Creating More Effective Graphs”. They are not as scary as Cleveland for non-statisticians.

“Creating More Effective Graphs” presents Cleveland’s ideas to a non-technical audience. Cleveland begins his preface by stating “This book is about graphing data in science and technology.” He suggested that I write a book applying these ideas to other fields.

I really like Doumont’s distillation of Tufte’s principles (http://www.treesmapsandtheorems.com/). He covers writing, presentation, and graphics all with the unifying motif “increase signal, decrease noise”.

Steve Kosslyn always has good ideas, though not always specifically on infographics. Yet, he’s quite accessible for the basics. I wish we had an update of Jacques Bertin, but it would be too complex for many. One of the most exciting finds is free on the web–watch Hans Rosling on TED on the web.

You could do far worse than the ggplot2 book/wiki!

I recommend two really awesome and very “visual” :-) books: Yau, N. (2011). Visaualize This. http://book.flowingdata.com/ and Ross Maciejewski, R (2011). Data Representations, Transformations, and Statistics for Visual Reasoning. http://www.amazon.es/Representations-Transformations-Statistics-Reasoning-Visualization/dp/1608456250

Not a textbook, but the materials for Stanford’s CS 448b may be helpful.

In addition to those already mentioned, Dona Wong’s book is a nice intro. And The Functional Art by Alberto Cairo, though there’s not much in the way of statistics in his book. I don’t think there’s a single source–Jeff Heer’s Stanford class notes has things pulled from lots of places.

I have Tukey’s EDA, Cleveland’s book and have long been a Tufte fan, but in fact those illustrate an issue that I hope some of the other sources address – thanks, I’ll look them up.

The issue is the combination of:

1) Black/white, greyscale, color

2) 2D vs 3D

3) static, vs dynamic vs dynamic-interactive

Some of the classic works of course were limited by the techno guy of the era, and by focus on print:

Black/white usually, 2D, static.

There are still roles for that, but the world has changed, when a smartphone has more 3D interactive graphics performance than high-end workstations 20 years ago, albeit on smaller screen.

Anyway, I hope somewhere there is a piece that helps students understand the available options, software tools, and the tradeoffs. (I still want better software for showing fuzzier density plots in place of lines and error bars :-))

I like M. H. Briscoe’s “Preparing Scientific Illustrations”. Although I sometimes disagree with the author, the material is clearly presented and the students get the point. I found some of the other books less appropriate for educational purposes. The Briscoe book deals with the fundamentals, such as eliminating clutter and choosing the right format (e.g., box plot vs line plot); it does not discuss technical details.

Cheers,

E.J.

Howard Wainer “A Trout in the Milk” has a lot of good stuff.

I second the recommendation of Stephen Few’s work. Good advice coupled with easy and understandable explanation. ‘Now you see it’ is his latest book.

For some of his ideas and articles check the website and blog: http://www.perceptualedge.com/

[…] Textbook for data visualization? by Andrew Gelman. […]

See Rafe Donahue’s notes on “Fundamental Statistical Concepts in Presenting Data: Principles for Constructing Better Graphics.”

I thought he does an excellent job of combining the statistical, aesthetic, and pragmatic aspects of data visualization:

http://biostat.mc.vanderbilt.edu/wiki/pub/Main/RafeDonahue/fscipdpfcbg_currentversion.pdf

Also, I second Alberto Cairo’s “The Functional Art” — it’s not heavy on the statistics, but it encourages readers to think deeply about their graphics, in ways that a statistician would approve of.

I’ve just realized that Andrew already highlighted Rafe’s notes on this blog, a few years earlier:

http://statmodeling.stat.columbia.edu/2009/02/12/fundamental_sta/