Visual Strategies – A Practical Guide to Graphics for Scientists and Engineers by Felice C. Frankel and Angela H. DePace, 2012

The production of a great figure is one of the most difficult jobs a scientist faces.  We are poorly trained in graphics and design, and we spend so much time creating the data, or the new understanding, that we loathe the time spend recreating it for an audience that didn’t share our path to that knowledge.  But a great figure, either in print or in a presentation, can be a world changing thing – it gives the viewer a rapid, inclusive grasp of the topic similar to what you have – and then enables them to follow you rapidly on the path of what we can do with that new understanding.  It is not so much that a great figure has more information in it, but rather that the information it has, is easy to assimilate and quickly generates an iconic understanding of the topic.  Rather than having to think about disjointed aspects of a problem, the viewer now has a single anchor point with many issues and related phenomena accessible from that point.

But how can we learn to create figures with this kind of impact?  Frankel and DePace bring their experience from photography and biology together, along with a bevy of extraordinary contributors, to give specific advice on how to great good figures, and most importantly, how to carefully edit and adjust those good figures to become great figures.  

This book is a story of before and after.  Almost every one of the examples is given in an original form, and then a greatly improved form.  In many of the case studies we see the figure as it evolves from the scientists first drawing, through multiple drafts, to a final version.  Felice and Angela give us several dozen examples where they show an original figure, and then analyze it according to five key steps to good graphics:

Compose  Abstract   Color   Layer   Refine

This list reflects the author’s experience from photography and biology, which share the importance of images in transmitting information.  However, in their extensive set of before and after figures, we see how these concepts are important to many kinds of figures, and how following these five steps can probably improve any scientific figure.  Their explicit description of which of the five steps were improved upon in each before and after set gives us a clear tutorial in the practice, and a beautiful set of examples for later consideration. Even the comparative statistical graph gets a twice-over with much improvement.Unlike some other graphics books, this book is purely for scientists.

Proportional diversity of species clusters through time.  Bush and Bambach (2010) Annual Reviews of Earth and Planetary Science. Each higher taxon within each cluster is colored according to its feeding mechanism, as indicated in the legend.  A subset of the data presented in the original figure is shown.

Proportional diversity of species clusters through time. Bush and Bambach (2010) Annual Reviews of Earth and Planetary Science. Each higher taxon within each cluster is colored according to its feeding mechanism, as indicated in the legend. A subset of the data presented in the original figure is shown.

The importance of color is repeatedly demonstrated – using consistent colors in being one of the key ways to help the viewer rapidly understand how items are related. I have found this to be one of the most difficult aspects of creating figures with multiple panels, such as related graphs.  Even though Powerpoint and graphing programs make it difficult to maintain that consistent color palette and relationship to data throughout multiple panels, it is extremely important to keeping the viewer oriented.  Their figure from page 23 is a great example of a complicated graphic, simplified enormously with consistent color.  (This is also an excellent example of the use of only data ink as we learn from Edward Tufte’s great books, with focused use of labeling and subtle lines to draw our attention to linked phenomena.)

Felice and Angela prove to us that the essence of a great figure is careful editing – which they call refinement. Two great examples of this are found in [page 41 and page 48] dealing with the importance of color. In the first case, a famous Frenkel improvement on a nearly monochrome original photograph by the investigator of a gold surface, patterned with hydrophobic lines which repel water.

Square drops of water from Whitesides et al. (1992) Science.

Square drops of water from Whitesides et al. (1992) Science.

Frankel's retake of the same phenomenon - with color added!

Frankel’s retake of the same phenomenon – with color added!

When water is added, a droplet pattern forms on the surface reflecting those lines. Felice’s version shows us that no matter how compelling the original, slightly-out of focus photo was, it is much more illuminating with color.  Adjusting the nature of the image to improve the information that it conveys is good science – it does not debauch the original experiment or result, but makes it much easier to appreciate. There is no harm in science being beautiful.

In a second example Frenkel and DePage go one step further – they hand color the original photographs in order to highlight the phenomena of interest, the movement of two particular bubbles.  By choosing slightly cartoonish colors on a black and white photo, and carefully stating in the caption that, “The addition of hand coloring follows two selected bubbles”, it is clear to the reader that this is an explanatory addition, and not data or analysis.  This type of addition to a photo must be done with full disclosure and respect for the photographic data, but doing so makes it possible to rapidly see the key point of the figure, while still using the original data image to make the point.

In my favorite before and after of the book, we see a classic circle of Powerpoint arrows – (which of us haven’t made this rough draft themselves!) turned in a much simpler graphic be recognizing that the topic is seasonal changes – and we all know the iconic order of the seasons.  Placed in simple circular form, the lack of arrows improves understanding, and makes room for more information on the graph. In the second, a set of molecular interactions with a surface was first described in a very compact, journal-format figure with panels for the interactions, and a panel showing the surface. In the revised version, which now takes up more space but is instantly understandable, the molecules are shown interacting in series with a long, continuous surface.  This immediately shows us what is the same, and what is different, for each of the five interactions described.  This approach turns the figure from something usable, but forgettable, in a journal, to something that would be instantly understandable in a visual presentation.  The desire to be incredibly concise, driven by our editor’s space limitations, often causes us to choose the dense format.  In the publication phase, we have to recognize that the expansion helps the reader, probably much more than would have the text we had to remove to make room for it. And in the presentation phase, the rapidly understandable figure is worth far more than the canonical thousand words.

But how do you create a great figure? Felice and Angela the concept of pentimento, literally regret, being the process by which artists change their works during creation.   In six stories from biology, astronomy, and statistics, contributing authors describe how they created their figures, including the intermediate versions.  I am particularly fond of their example of the imaging of dark matter using the Hubble Telescope.  The final version contains real data, an iconic form of the telescope itself and its orientation toward the imaged fields, and finally a 3-D rendering of the data, with a detailed description of how the rendering was done, and some discarded versions of that process.

The figures in this book all represent incredible labors, many hours or even hundreds of hours often from teams of scientists and artists.  What can every scientist extract from these awesome examples?  Let me suggest two items. First, great figures take time.  Second, in order to be understandable, great science has to be easy to consume, even for the expert reader. The time that you take preparing an elegant figure is repaid many times by the ability of the viewer to rapidly comprehend your science.  Championing science is about transmitting knowledge and understanding.  The time you save your audience gives them time to grasp and evaluate your science. They then themselves multiply that time into new understanding and new applications, because they get it, they are excited, and you have taken the time to show them how it matters.

Felice and Angela maintain an informative blog on this topic at www. visual-strategies.com.