Seeing Beyond the Data: Lessons from Faith, Text, and Visual Representation
My synagogue recently adopted a new prayer book for the High Holidays, the period of the Jewish year dedicated to reflection, repentance, and renewal. What caught my eye in this new book were its footnotes translating certain prayers or describing the origin of others. During this year’s services, I found one footnote in particular that inspired me to reflect not only on my religion but also on my day job of creating visual content to communicate data, research, and analysis.
In the Jewish tradition, we cannot worship other Gods nor represent God visually, say in a sculpture or painting. The note that caught my eye interrogates this idea:
But why is a visual representation of God inappropriate even though a verbal representation is appropriate? A picture is meant to capture the entire essence of what is being represented; it strives to create a full representation, leaving no gaps. Not so language: a verbal description is only partial, and the open spaces it leaves make language an appropriate medium for representing God. Representation in a visual medium, meanwhile, diminishes God and thereby desecrates the holy….
-Moshe Halbertal, Of Pictures and Words: Visual and Verbal Representation of God in The Divine Image: Depicting God in Jewish and Israeli Art
In Halbertal’s argument, language only gives us only a partial description of God. Art, by contrast, is too fixed, too definite, too concrete. It tells us too much about God and because of that, it is apt to be wrong.
I’ve been thinking about how these concepts extend to our work visualizing data. If visual representation is insufficient, can text effectively make up for the difference? Or is the practice of data visualization inherently incomplete?
Take a straightforward chart, like a bar chart. The bars in that chart summarize an entire data set down to a single set of numbers. The data themselves summarize the experience of the people or firms or countries. The data we use in our charts are not complete—they miss details and nuances and subtleties that the original surveys or forms are not going to capture. We then take these incomplete data and transform them into averages or sums in a rectangular shape.
From collection to visualization, our data are always incomplete. The representation, as Haberstal points out, leaves gaps.
With this incompleteness in mind, then, how can we use text to fill in the gaps? Good, active titles are a good starting point. Borkin and colleagues (2015) demonstrated that people do read titles, and Kong and colleagues (2018) find that bias in chart titles affects reader’s perceptions of the data.
Labels and annotations are another way to fill in the gaps. Sometimes, we need annotations to help explain to readers how to read the graph, and other times we can use labels to explain the content and key findings. In either case, text inside the graph space can help add context or a broader perspective to the data.
The Habertal quote continues:
Walter Benjamin has argued, famously, that the aura of representation has been lost in the era of reproduction. An abundance of pictures and excessive visual presence can undermine the sublimity of that which is represented. The absolute prohibition on visual representation forestalls the harm that Benjamin saw in the connection between representation and desecration of the holy.
The 1935(!) Walter Benjamin article that Haberstal cites explores how the mass reproduction of art—through photography, film, and other technologies—changed the nature of art and its role in society.
Today, technology has clearly changed how we produce and reproduce data and data visualizations, which have also changed its role in society. We’ve moved from hand drawn graphs and charts to GUI-based tools like Lotus Notes, then Excel, then faster and better programming languages like R, D3, Python, Svelte, and more.
I’ve been exploring how going back to physical data projects and exhibits can engage and inform. That practice (and growing academic literature) moves us to interact with data in different ways that may be more limited in their scope or reach but may also be more useful or interactive for some participants.
I’m not arguing that we should all reach to our religious texts to find guidance on how to make our next dashboard or graph. But we should be careful about how data visualization can—correctly or incorrectly—suggest a full explanation of some finding or pattern that may need additional caveats, instruction, or discussion.
Podcast: Mastering the Art of Executive Presence with Dr. Laura Sicola
Dr. Laura Sikola, a cognitive linguist and communications expert, visits the podcast to talk about how to be a better presenter and communicator. Laura’s work focuses on how to have better presence, telling stories, and connecting with your audiences. We talk about communication challenges like the “expert’s curse” and how developing your “executive presence” can command attention and build rapport.
Things I’m Reading
Books
Articles'
Frames and Slants in Titles of Visualizations on Controversial Topics [re-reading] by Kong, Liu, and Karahalios
A Qualitative Analysis of Common Practices in Annotations: A Taxonomy and Design Space by Rahman et al.
TV & Movies
The Diplomat, Netflix
Silo, AppleTV+
Shrinking, AppleTV+
Note: As an Amazon Associate I earn from qualifying purchases.
Supported by: Curiosity Box
The Curiosity Box is the world's first subscription for thinkers. Premium science toys, experiments, and collectibles designed for curious adults & delivered to your door
4 times per year.
Supported by: Automattic
WordPress Hosting That Means Business. Fully managed from top to bottom, our platform delivers incredible stability, tooling, and performance for your most important sites. Everything you need to get your site running.
