What makes a visualization good?

"I think some people hold to a myth of an objectively good visualization. But I prefer to think about it in pragmatic terms. So a visualization is good if it lets us discover, confirm, or communicate something important about our data."

—Jeff Heer

Recently, the team over at Sphere had the chance to sit down with Jeff Heer, co-collaborator on data visualization tools like D3.js, Vega, and Vega-Lite.

In that discussion, Jeff addresses the question, "What makes a visualization good?"

If you have experience working with data, you probably know the most common answer: to help you quickly make sense of the data, so you can better understand, strategize, and act on it. 

But Jeff’s answer had more nuance.

In short, Jeff pointed out that the idea of a “good visualization” is entirely dependent on your initial goal for collecting that data. Sometimes, you need information conveyed quickly, and sometimes you need it conveyed deeply. 

Again, what makes a visualization “good” always comes back to the question, “What’s your goal?” 

Check out Jeff’s initial answer, and click here to see his full response (& you’ll also be able to learn more about his upcoming course on data visualization).