Mark Baillie, Alison Margolskee, Baldur Magnusson, Andrew Wright, Ruquan You, Ivan-Toma Vranesic, Julie Jones, Marc Vandemeulebroecke
Novartis
Objectives: The goal of this work was to compile a set of Good Graphical Principles concisely and tangibly for everyday use in the production of scientific graphics. Good graphics are at the core of exploring and understanding data, communicating results and conclusions, and ultimately are critical to support decision-making. Hence, due attention should be paid to doing them well – and to understanding what “well” means in the first place. This is the purpose of the Graphical Principles Cheat Sheet.
Methods: A large body of literature and training material, including Tukey (1977), Cleveland (1985), Tufte (2001), Doumont (2009), Few (2012), Krause and OConnell (2012), Robbins (2013), Wong (2013), and Duke et al. (2015), has been condensed into an easy-to-use single-page reference sheet. The various principles have been grouped into sections such as: selecting the right base graph; an effectiveness ranking of graphical attributes (volume, color hue, depth, area, angle, length, position, etc.); how to facilitate comparisons; the best use of color; enhancing legibility and clarity; various implementation considerations; and a checklist for users to assess a graph againstthe most important of these aspects. Each point is illustrated concisely and intuitively with a thumbnail graph. Key references and online resources are included on the sheet as well. Unlike other, more general graphical reference sheets that are available in the world wide web, our Cheat Sheet is tailored for the practicing modeler and statistician in biopharmaceutical sciences and drug development.
The Cheat Sheet has been widely distributed to associates within Novartis as hard copies at a launch event and in electronic format as part of an internal gallery of good graphic examples plus associated code. Workshops have been run for smaller groups of associates, focused on working through an example of how to improve a graph with reference to the points made in the Cheat Sheet. This series of workshops is a key step in ensuring take-up of the messages in the Cheat Sheet and reinforces the idea that the Cheat Sheet is a truly practical tool for improving graphics.
Results: We present and share a carefully designed “Graphical Principles Cheat Sheet” for everyday use in graphical data exploration and the production of graphics for communicating analysis results and conclusions. Our poster re-formats the Cheat Sheet to a large scale, and the Cheat Sheet itself is available as a hardcopy hand-out with the poster. Modelers and statisticians at Novartis now use this Cheat Sheet successfully in their daily job.
Conclusions: A carefully designed single-page reference sheet on Good Graphical Principles is a very useful tool for the creation of clear and impactful graphics. We have had good experiences with our Cheat Sheet at Novartis and now want to share this piece of work for the benefit of a wider audience. We would be interested in an open exchange with other interested parties that may pursue similar goals.
References:
[1] Cleveland, WS (1985). The elements of graphing data. New York: Chapman and Hall.
[2] Doumont, J (2009). Trees, maps, and theorems: Effective communication for rational minds. Brussels: Principiae
[3] Duke S, Bancken F, Crowe B, Soukup M, Botsis T, Forshee R (2015). Seeing is believing: Good graphic design principles for medical research. Statistics in Medicine 34 (22), 3040-3059.
[4] Few, S (2012). Show me the numbers ‐ designing tables and graphs to enlighten (2nd. Edition). Burlingame, CA: Analytics Press
[5] Krause A, OConnell M (editors, 2012): A picture is worth a thousand tables. New York: Springer.
[6] Robbins NB (2013). Creating more effective graphs. Chart House.
[7] Tufte ER (2001). The visual display of quantitative information, 2nd ed. Cheshire, CT: Graphics Press.
[8] Tukey J (1977). Exploratory data analysis. Reading, MA: Addison-Wesley.
[9] Wong DM (2013). The Wall Street Journal Guide to Information Graphics: The Dos and Don’ts of Presenting Data, Facts, and Figures. WW Norton.
Reference: PAGE 27 (2018) Abstr 8416 [www.page-meeting.org/?abstract=8416]
Poster: Methodology - Other topics