Effective Data Visualization
The MCAA has initiated a series of webinars that run each second Friday of the month starting from December 2021. The aim is to provide educational webinars to the MCAA members and non-members as a way to thank for the support received all these years. The webinars are co-organised by the Communication WG in collaboration with a different MCAA Chapter or Working Group each month and cover a variety of topics.
In the fifth episode, organized by the Communication WG, the focus was set on a very important topic: Effective Data Visualization.
Great data visualization pops. Our information-hungry brains dive in and swim, proving the claim “information is beautiful” and yet, the attractive colour schemes, elegant designs, and cascading interplay between numbers, measurements, positions, and shapes can sometimes belie the dire, life-and-death information contained therein. In short, good data visualization attracts the eye and tells a story; it then helps us move from the realms of aesthetics and storytelling into clear and informed action.
Here we will summarize the highlights of the 5th Episode of the MCAA Around the World Webinar Series, which can also be watched on YouTube.
Striking this balance is the purview of scientist-turned-data visualization expert Cédric Scherer who offered an engaging talk on “Effective Data Visualization” in the latest Marie Curie Alumni Association’s Around the World Webinar series.
Cédric spoke about his shift from earning a doctoral degree in ecology to excelling as an independent data visualization specialist. While his professional skills are displayed by his recent visualization for Scientific American showing the escalating drought conditions of the Western United States, he also applies his talents to social issues, creating graphs like this to show the persistent salary gap in Europe:
For the webinar, Cédric began with a short history of data visualizations, including Florence Nightingale’s famous rose chart of mortality rates and causes for British troops during the Crimean War — catch up with one of our previous blog post on the topic (find it here).
After establishing some of the tenets of the practice, Cédric expanded upon four qualities that make for good data visualizations:
Referring to the integrity of information, he reminds us that, while we must work to be as exact as possible, “data is never a perfect reflection of the real world” as we are often working with subsets of data made available by humans often working with machines. As for storytelling, while we often have the urge to craft data to support a thesis, its often more impactful if we used data to “teach us what isn’t true.” Of course, the choices one makes during data visualization process can support specific narratives or goals, as the side-by-side comparison of “illegal immigration” and “annual migration” graphs show.
While some scientists may not immediately notice the subtle shifts from arrows to dots or red to blue, the comparison reveals how visualizations can draw upon our audience’s implicit biases.
For those seeking clear and direct advice, you’ll hear a convincing argument about why we should never use word clouds.
While word clouds have are an easy solution that many of us turn to when pressed for time and ideas (raises hand + raise hand emoji), alternatives like bar charts, network diagrams, and the almost always appealing Sankey Diagrams can offer clear and accurate insights.
In the latter half of the webinar, Cédric reviews some inspiring possibilities for encoding multiple dimensions, including dumbbell charts, decision trees and my personal favourite, this “dot plot merged with interval stripes” to show ratings for coffee beans around the world.
Finally, Cédric offers us a crash course on colour and colour palettes. Of course, colours elicit emotional responses, and many of us remember refreshing the John Hopkin’s Global Covid-19 Dashboard with a tinge of horror, as the blood red dots expanded and seemed to swallow states and nations in a dismal sea of helplessness. While the red alerts can be useful, as proven by Ed Hawkin’s “warming stripes” which show historic rise of global temperatures, different types of colour palettes can be selected based on the types of data to be displayed.
Data visualization offers researchers the opportunity to grab their audience’s attention; watch the full webinar on YouTube and visit Cédric’s webpage to learn more about how to make dataviz pop.
Author
Daniel Wuebben, Adjunct Professor of Liberal Arts at NYU-Madrid and Communication Working Group Manager of ISGAN
