Sketchy Science

You can’t see every item in a block of text. See every item. Force the insight.

Mar 28, 2023

Welcome to Chartography: insights and delights from the world of data storytelling.

This week we begin with two historic delights before proceeding to the latest installment of How to Value Data Graphics. Enjoy!

Sundries

🎨 This German map uses color to plot the value of farmland, from yellow (least valuable) to blue (most).

Its bivariate color palette uses fill pattern/texture to indicate another dimension: average farm size in each district, identified horizontally in the map’s color legend.

Seeing this map, cartographer and author Kenneth Field observed, “Ordinarily you'd recommend no more than 3 classes (9 colours) for the two variables on a bivariate choropleth, but introduce patterns and you can get a 5x6 matrix and still distinguish 30 separate combinations.”

Explore over one hundred maps from the Deutscher Landwirtschaftsatlas (Berlin, 1934) in the David Rumsey Map Collection. link

💰💀 One of my favorite pictorial unit-charts flipped a money bag upside-down to make a graphic connection between war profiteering and deaths. Its accompanying text began by stating that "the desire for profits makes men do even stranger things than destroy goods." I recently digitized the design:

Learn more about the 1935 original and see the full re-presentation at infoWeTrust.

How to Value Data Graphics, Part 6

We are in the midst of exploring the value of reading charts. These middle topics include the most commonly cited reasons why data visualization is useful.

Last newsletter discussed inspecting machines and exploring data. See that essay (Inspector Explorer) for an overview of the entire How to Value Data Graphics series.

Today, let’s discuss the relationship between data graphics and insight. But first, a reminder that the foundational observation that set this series in motion is that insight is overrated (Beyond Insight). I have found that insight is wonderful when it happens, but is not necessary for a graphic to be valuable.

Discover Insights

The history of science is littered with visual tools that helped develop breakthrough ideas. DNA’s double helix, the periodic table of elements, the theory of relativity, and the theory of evolution all credit visual thinking in some way.

There are fewer famous stories about a single graphic directly producing a critical eureka! moment. The canonical example of a graphic discovery is Francis Galton constructing early weather maps and finding the circulation of wind around a pressure center.

The anticyclone, diagrammed below, is a phenomena that cannot be detected from a single weather station. You must collect, assemble, and plot data from across a continent to see the pattern.

Today, graphics persist throughout science as helpful tools, such as Nextstrain’s colorful real-time tracking of pathogen evolution, one of the centers of scientific discourse throughout the Covid-19 pandemic.

There are many ways to make graphic discoveries. Tukey said it best in the canonical soundbite:

The greatest value of a picture is when it forces us to notice what we never expected to see.
—Original emphasis from John Tukey, Exploratory Data Analysis (1977), p. vi.

Bedtime hero-stories we share about how discoveries happened usually compress years of iteration, approaches, and collaboration into singular moments. In reality, historically and today, making discoveries with graphics is a messy affair. Sometimes, we can distinguish the graphic for giving the extra oomph that pushed everything forward.

Making monumental discoveries is complex. But teaching them is a more elegant affair. Data graphics excel at fostering personal discoveries. These are the learnings that you help your audience see. As a data storyteller, I consider Tukey’s quote as a kind of target for my work. A picture can show, force, a new idea with clarity unlike other media.

Below, see Hans Rosling presenting one of his public health visualizations. It shows the extraordinary progress of life expectancy in Vietnam (red) across a few decades, against the context of leftward falling fertility rate—and compared the United States (yellow).

This static image suffices, but his live presentation is undeniable. In addition to his enthusiastic character, the graphic’s comparisons—between axis dimensions, over time, and between nations—is what makes this presentation soar.

For the past year I’ve lectured about the power of asking “compared to what?”: To help audiences engage and find meaning, teach insights via easy-to-make comparisons. Present them with comparisons they make automatically. Present them with comparisons that are undeniable. For example . . .

▴ ▲

. . . Everyone saw that one triangle is bigger than the other, whether they wanted to or not. (Constantly asking “compared to what?” is also a simple and effective way to navigate your own information consumption.)

Another modern example of a graphic-powered insight comes by way of the Wall Street Journal. It is from a series of before/after comparisons that illustrated the effectiveness of vaccines with dazzling mosaics. Here’s the measles variant from the series:

The bottomline insight of both Rosling’s Vietnam and WSJ’s vaccination could be transmitted with a short sentence. But neither text would achieve what Tukey described. They would not force you to engage with the insight.

We could have just Vietnam’s beginning and end (1965 and 2007), but that would not be enough to force the insight. Seeing every year matters. And you can’t see every year if it is a block of text.

We could have just a national metric of disease incidence for before and after vaccination, but that would not be enough to force the insight. Seeing every state matters. And you can’t see every state in a block of text.

Going forward

Seeing every item fosters trust with the key insight. It’s easy to fake directions to a location, it’s hard to fake a map. Seeing every items also affords comparisons between items—it helps us appreciate the landscape context. And that is what we will explore next newsletter.

Until then. Onward!—RJ

Notes

Follow Kenneth Field on Twitter. link
Sir Francis Galton produced an extraordinary number of graphic innovations, perhaps more impactful than anyone after William Playfair. Wikipedia
“Nextstrain is an open-source project to harness the scientific and public health potential of pathogen genome data. We provide a continually-updated view of publicly available data alongside powerful analytic and visualization tools for use by the community. Our goal is to aid epidemiological understanding and improve outbreak response.” link
Read John Tukey’s Exploratory Data Analysis from the Internet Archive. link
Watch all of Hans Rosling’s TED talks. link
Tynan DeBold and Dov Friedman, “Battling Infectious Diseases in the 20th Century: The Impact of Vaccines,” The Wall Street Journal, 11 February 2015. link

About

Data storyteller RJ Andrews helps organizations solve high-stakes problems by using visual metaphors and information graphics: charts, diagrams, and maps. His passion is studying the history of information graphics to discover design insights. See more at infoWeTrust.com.

RJ’s recently published series, Information Graphic Visionaries, a new book series celebrating three spectacular data visualization creators. With new writing, complete visual catalogs, and discoveries never seen by the public. His first book is Info We Trust, How to Inspire the World with Data.