A Dangling Blue
Something needs to change with how we craft charts.
Welcome to Chartography.net — insights and delights from the world of data storytelling.
🍎 I’m returning to Cooper Union to teach data graphics across a five-session virtual course. Registration just opened: https://coopertype.org/events/principles-of-data-graphics-2
🍎 My office hours are available on Tuesdays and Thursdays. Sign-up to chat charts: https://calendly.com/infowetrust/office-hours
This edition marks one of the ideas I’m taking with me into 2025. I’m only able to partially articulate it, but that doesn’t mean I don’t believe in it. You may want to take this idea with you too.
Let’s start with a car metaphor.
When driving a car through a turn you don’t feed the machine a series of precise angles using a number pad.
Instead, you look where you want to go and let the visual feedback guide your hands through the curve.
The angles of the turn are an unnecessary technical abstraction. They theoretically exist, but neither you nor the car want them.
The car’s interface, its steering wheel, is an amazing natural linkage between the goals of a human mind and a powerful machine. No special training is required to steer a vehicle. A toddler can learn to drive a battery-powered car given a full charge and an empty driveway.
When I make a chart, I rarely feel like I’m effortlessly steering the machine to my goals. Despite data graphics being a communication art, their construction is still about inhuman inputs to machines. Try designing a color palette. Try doing anything fun with circular diagrams. Try doing anything creative with any chart.
There is a faint sense that something needs to change with how we make digital charts. I believe we are doing it wrong.
Maybe it’s a faint hope.
Let’s have an example closer to home. Take color. “Bluebird” was Krylon’s color of 2024.
Consider a handful of the many ways we could describe “Bluebird” (if you are familiar at all with color science then the first three should be familiar):
1️⃣ The hexadecimal code for Bluebird is #2e9bd3. Hex codes are good at compressing the recipe for the right amount of red (2e), green (9b), and blue light (d3) needed to create Bluebird on a digital screen.
Base-16 hex codes use the same six spaces to represent far more possible combinations. (16^6 vs 10^6, or if you prefer, 256^3 vs 100^3.) However, because we do not have single characters to represent numbers larger than 9, we have to express hex with letters: 0-1-2-3-4-5-6-7-8-9-a-b-c-d-e.
Who reads hex codes and sees the color in their head? You aren’t supposed to. Hex is a compressed recipe for transmitting colors efficiently.
2️⃣ Human-readable hex codes is RGB (red green blue). RGB converts Bluebird from its base-16 hex values to familiar base-10 numbers: red 46, green 155, blue 211.
Access to the recipe for how bright to make individual LEDs does not help us see Bluebird in our head. We aren’t naturally talented at mentally mixing light primaries to get the colors we want. We also don’t naturally think about numbers on a scale from 0 to 255, we like 1 to 100.
RGB is a readable recipe, but not much more useful to the color-designer than hex. Recipes are for telling screens how to dial up and down LEDs. This is an odd way to think about color. I’ve never increased or decreased a particular RGB value to get closer to what I want.
3️⃣ The hue-saturation-brightness (HSB) color system can help us tweak our designs. In HSB, Bluebird has a hue of 200°, saturation of 78%, and brightness of 83%.
We can guess how Bluebird will change with different saturation or brightness. Imagine lowering to 50% saturation—it would look more dull. Imagine increasing to 95% brightness—it would look brighter! Even better, these values are expressed on a 100-point scale, allowing us to easily program something that is, for example, half-saturated.
HSB comes stock on most digital tools. I use it to build noticeable gaps in brightness between my colors, then tweak saturation after brightness is set.
HSB helps change colors in relative ways, but it isn’t perfect. HSB is tuned to how the computer produces color, not how people perceive color. Colors with noticeably different brightness, for example, may share the same exact same brightness-value.
4️⃣ Krylon didn’t talk about any of this when promoting Bluebird. Here’s a sample of how they marketed their color of the year:
Refreshing contentment. Bluebird both uplifts and comforts.
This bold, yet pale-hued pastel creates a connection with blue spaces like those found in nature, harkening to feelings of floating in soothing ocean waves.
Bluebird evokes joy and contentment within a space, embracing consumer desire for ‘dopamine decor’.
Easy. Breezy. Modern minimalism takes flight as Bluebird spreads the beautiful colors of oceans and lakes to create a calm, uplifting effect.
Acknowledging that these Krylon blurbs are marketing-schlock, I want us to appreciate that this copy is completely focused on the human side of color. It isn’t concerned with how to formulate color, it cares about how color effects people, how color creates meaning.
The systems we use to make charts are in service to the machine, not to us. We make charts in terms of fixing data dimensions to encoding channels. Year on the horizontal. But you do not want your audiences talking about your work in those terms. You don’t want anyone thinking in terms of encoding channels.
Digital charting tools are largely based on ideas presented in Leland Wilkinson’s Grammar of Graphics (1999). I celebrate that the grammar is there, but I am agitated that fluency in our craft is predicated on knowing its grammar. Plenty of beautiful and effective communicators do not have any technical understanding of grammar. Why shouldn’t it be the same with charts?
You don’t want any audience thinking in terms of encoding channels. What you want is to attract their attention to something worthwhile, you want to force them to notice what they never expected, you want to open their eyes—everyone’s eyes—a little wider to the reality swirling around us all.
The gap between Bluebird’s hex code and its marketing copy is wide. The gap between how we make charts and how we consume charts is similarly wide.
When I make a chart I imagine myself a creator, maybe even a performer. I wish to feel like a singer in front of a microphone. But I am more often like the sound engineer behind a mixing board, twiddling dials on a machine and inventing new technical gimmicks to achieve what I want.
One of the epigraphs in the Grammar of Graphics is that “Bad programmers ignore details. Bad designers get lost in details.” Today, is it possible to design a chart without getting lost in the details? I fear not.
I’m only able to partially articulate this idea, but that doesn’t mean I don’t believe in it. Call it an artist’s intuition or a fool’s confidence. I appreciate you helping me think it through.
Onward!—RJ
About
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 next book, Info We Trust, is currently available for pre-order. He published Information Graphic Visionaries, a book series celebrating three spectacular data visualization creators in 2022 with new writing, complete visual catalogs, and discoveries never seen by the public.
Great edition RJ, thanks for articulating this idea as best as you could!
I have had a similar hunch in the past and you've defined it better than I could.
I have a feeling that focusing on encodings or data treatment to explain a chart only further exacerbates the gap between designer and audience. Most people have no interest in visualization encodings whatsoever, they want to feel that they understand the chart's argument and can employ it in conversation.
I love this metaphor of chart designers as singers in a band. It leaves me thinking that we may be forcing a too-rigid distinction between visualization as a technical tool for creating knowledge and a means for self expression. Maybe they are one and the same, and we should find ways to treat it this way.
Thanks again for these reflections and happy 2025.
Have a nice one!
Two observations:
1. There’s an old anecdote about when artists gather, they don’t talk about “art” but about the best places to buy paints and brushes.
2.Apple originally bridged this gap in the Macintosh with apps like MacPaint. Perhaps what’s needed is more intuitive interfaces ala Kai’s Tools for Photoshop (r.i.p.).