Friday, March 22, 2019

Edward Tufte and Drawing Figures

Ten years ago, I wrote a blog post about Edward Tufte's book The Visual Display of Quantitative Information. Recently, we discussed this book in a graduate class I am teaching. One goal of the course is for students to learn how to write scientific papers, including drawing figures.

I like Tufte's advice (below) about friendly data graphics. I agree with everything he says, except for his distaste of sans serif fonts.

Advice about making friendly data graphics. From The Visual Display of Quantitative Information, By Edward Tufte.
Advice about making friendly data graphics,
from The Visual Display of Quantitative Information,
by Edward Tufte.
I wanted to give the students examples of how to improve illustrations but I didn't want to pick on a colleague, so I found a couple figures from my own papers that could be friendlier. The first one is from an article about magnetic stimulation that Peter Basser and I wrote. On the left is the original, and on the right is my revision.
Figure 3c from Roth and Basser (1990) A model for the stimulation of a nerve fiber by electromagnetic induction. IEEE Trans Biomed Eng 37:588-597.
Figure 3c from Roth and Basser (1990) A model for the stimulation of a nerve fiber
by electromagnetic induction
. IEEE Trans Biomed Eng 37:588-597.
The purpose of this figure was to show where magnetic stimulation occurs. That message was in the original figure, but you had to read the caption carefully to find it. In my revision, I clearly marked the locations where depolarization (that is, excitation) and hyperpolarization occur. The big square surrounding the original figure is what Tufte would call "chartjunk" and I deleted it. Instead, I tried to focus on the data. I also labeled the nerve and the coil, so you don't have to read the figure caption to determine what's what. Including units on one of the numerical values practically eliminated the need for a caption at all. I confess, the original figure was cropped from a mediocre scan of the article. Therefore, let's not focus on which figure is crisper, but rather on the overall design. Also, the 30-year-old original data is long lost so I had to retrace the contours in powerpoint using the polygon tool. If you look at the revised figure using high magnification, you may be able to see this. Nevertheless, in my opinion, the revised figure is better.

Another example is from a paper about the electrical stimulation of cardiac tissue.
Figure 3f from Sepulveda, Roth and Wikswo (1989) Current injection into a two-dimensional anisotropic bidomain. Biophys J 55:987-999.
Figure 3f from Sepulveda, Roth and Wikswo (1989) Current injection into a
two-dimensional anisotropic bidomain
. Biophys J 55:987-999.
The message of this figure is that adjacent regions of depolarization and hyperpolarization form when tissue is stimulated by a cathode. By shading the hyperpolarized region, I emphasized this message. I indicated the fiber direction, which is crucial to the main conclusion (tissue hyperpolarizes along the fiber direction). I eliminated the outer circle and subdued the coordinate axes to highlight the data, and inserted a black dot at the location of the cathode. You can decide if it's an improvement. A different version using color and containing all four quadrants is shown below.
A color, four-quadrant version of Figure 3f from Sepulveda, Roth and Wikswo (1989) Current injection into a two-dimensional anisotropic bidomain. Biophys J 55:987-999.
A color version of Fig. 3f from
Sepulveda, Roth and Wikswo (1989).
Another confession: each original illustration was one frame from a multipanel figure. They might have been drawn differently were they stand-alone figures (but I doubt it).

Five good books: The Visual Display of Quantitative Information, Visual Explanations, Evnisioning Information, and Beautiful Evidence (all by Edward Tufte) and Intermediate Physics for Medicine and Biology.
Five good books,
four by Edward Tufte.
Tufte has published several books on visualizing information, but my favorite remains The Visual Display of Quantitative Information. You decide if Russ Hobbie and I follow his advice in Intermediate Physics for Medicine and Biology. Sign up for his course on presenting data and information at his website.

Listen to Tufte talk about the future of data analysis in the video below.

Edward Tufte talking at the Machine Learning and Data Science Summit 2016 Keynote Session.

No comments:

Post a Comment