Humans are visual creatures, and today, when the national conversation is framed by data-driven topics like climate change, data security, and gerrymandering, data visualization has never been more important. Yet for designers and engineers, a palette, arguably the most impactful design choice when it comes to data viz, can be difficult to get right.
Susie Lu and Elijah Meeks are both senior data visualization engineers at Netflix, where they work on internal visualization products, but they recently made a tool they built independently available to anyone. Borne out of what Lu describes as frustration with existing color-picking tools, their web app Viz Palette, uncovered recently by Flowing Data, lets you test your color palette for legibility in different formats, line weights, backgrounds, and font colors. It also shows you what your colors look like to users with different types of color blindness, like deuteranomaly and protanopia. If two colors are too similar, Viz Palette lets you know–and allows you to randomize the sample data to see it with fresh eyes.
The duo’s frustration with color picking reflects the fact that color is a shifty, mercurial thing. Two colors that look different to me might look the same to you. A shade seen on paper might look different on a phone screen. Background color and other factors could totally change the way data is perceived. Even using existing scientific palettes–like the dreaded rainbow–can have its own pitfalls.
“I think most people assume that the traditional data visualization palettes are inherently better when they’re really just established, and their conservative dullness is unintentionally imparted on the work they’re used to create,” Meeks says over email. “Most people don’t have 20 distinct dimensions, or even 10, yet they’re using palettes that are optimized for them and as a result making their work look as indistinguishable as possible because of some positivist superstition that color is hard and that they don’t want to mess up picking a palette.”
It’s a useful addition to color tools that don’t go quite as far with their analysis, and Lu and Meeks are using it to put together an internal style guide for Netflix. It could also be useful for anyone working with color, whether in industrial design or app development. Products both physical and digital can have untold effects on the world around us, and tools like this one help bring some empirical evidence to the colors we give them.