Data visualization is like writing…
In high school, I learned the rules for writing an essay. You know them: Start with an introduction, then move onto your body paragraphs (each of which should contain at least two supporting facts), end with a conclusion, and, whatever you do, never ever say “I”. Throughout high school, I stuck to those rules and never strayed from them. Fortunately, the formula seemed to work for me; my writing was good and I did well on my essays. But when I arrived at college, I started reading pretty much anything I could get my hands on — classic novels, philosophy, news articles, non-fiction, etc. — and I noticed something: they often broke the rules. It was shocking at first. But I started to realize that, at times, it was okay to break these rules. I even started to slowly introduce such rule breaking into my own writing. And what happened? My writing improved. I had more freedom. Unconstrained by the rules, I found my own voice.
My writing would have always been merely mediocre unless I had started to explore those boundaries.
I remember taking a religious studies course and we were asked to write an essay on Shadowlands, a movie about C.S. Lewis, the famous Christian theologian and author of The Chronicles of Narnia. In the film, Lewis’s life was largely devoid of any real passion until he met Joy Davidman, a brilliant poet from the United States. As I watched this movie, I couldn’t help but reflect on how it made me feel, how it made me think about life and its meaning. So, when I wrote the essay, I broke rule # 1 — I used “I”. And I did so repeatedly. Despite breaking this cardinal rule, my professor loved it. She asked me to read it in front of class and, as I continuously used the word “I”, there were audible gasps from my fellow students. This was simply going too far. When the professor asked me if I had any advice for my fellow students, I told them “you don’t always have to follow the rules.”
In high school, I learned the rules for writing essays. Having mastered them, I realized that they were not so much rules as guideposts. Knowing this, I began to explore and experiment with the boundaries and gained a better understanding of when it was okay to wander outside of them. At times, I wandered when I shouldn’t have. Other times, I wandered too far. But those mistakes were just part of the journey. My writing would have always been merely mediocre unless I had started to explore those boundaries.
Data visualization is like poker…
Fast forward about 10 years. No Limit Texas Holdem Poker had become very popular in the United States — online poker sites were popping up every day, and there were no less than a dozen different poker-related television shows. I watched it on TV whenever I had a chance and, though rare, I relished any opportunity I had to play for real with family or friends. But, when I was playing, I’d often make bold moves. I’d go all-in on a terrible hand, hoping to bluff everyone out of the pot, or I’d slow-play a good hand in an attempt to drain more money out of my opponents. Why? Because that’s what the pros did on TV, so that’s what I thought it took to win. But, you know what? I never won. In fact, I was almost always one of the first to be eliminated in a tournament because of my overly-aggressive play.
The problem was that I never really learned the basic rules of poker. Instead, I thought that watching a one-hour television show was all I needed. What I didn’t realize is that these television shows were condensing hours-long tournaments into bite-size nuggets. And, for these shows to be entertaining, they couldn’t just show people checking all the time. So, they focused on the bold moves. Unfortunately for me, I came to the (wrong) conclusion that, to be good at Texas Holdem, you have to be bold and aggressive all the time. But the fact is that those actual bold moves are few and far between. Most of the time, the pros follow the guideposts closely, making the “right” play for their hand, the size of their chip stack, the number of players in the hand, etc. Every once in a while, though — when the time is right — they make a move and break the rules.
What does this have to do with data visualization?
These experiences shaped how I feel about the “rules” of data visualization (i.e. “Best Practices”). They are there to help ensure we don’t go astray, but they are not unbreakable laws, and should not be seen or followed as such. Let’s take bar charts, for example. There is a reason they are one of the most commonly used charts. They are scientifically proven to be one of the most effective types of charts in our repertoire. But this science only holds up when the bar chart is created in a horizontal or vertical fashion. What if I were to wrap the bar chart around a center point to create a radial bar chart?
Is this a good way to represent this data? Many of us would scream “NO!!!” but that would be a mistake without first understanding the data. Context is critical when applying best practices, so let’s add some.
Are we still screaming no? With added context, we can see that the radial positioning of each bar represents the actual direction of wind. So, the point of this visual is not simply aesthetic (though aesthetics are important, and I’ll get to that later).
Of course, this radial bar chart does reduce our ability to understand precise differences between one bar and the next. To mitigate this somewhat, the chart uses a background grid to help the viewer to compare lengths. But I’d also argue that the ability to see exact differences in individual bars is not terribly critical in this use case. Take, for example, the two highlighted bars below. Can you tell the which one is longer?
Probably not. With a regular bar chart, you could, though.
But does our ability to distinguish between the lengths of these bars matter in this case? Not really. The point of this high-level chart is to provide the viewer with an overall understanding of wind patterns at the Philly airport. And it does that very effectively, providing some instant insights, such as the fact that there isn’t a lot of southeasterly wind. So, in this case, the need for exact precision is outweighed by the visual insight provided by the radial layout of the bars. In short, this is an example where it’s okay to break the rules.
But, like Texas Holdem, we must also realize that most of the pros follow the rules most of the time, only breaking them when they feel it’s necessary and only after understanding the reasons and science behind these best practices. This is often obscured within our community, however, because so many of us (in our own personal projects) break these best practices, either to get noticed, to learn new skills, or to build something that is really beautiful and creative. In some ways, much of what we create might be considered more art that data viz (that’s a topic for another time). So, it’s logical that many new people could, after observing our work, come to similar conclusions about data visualization — the pros always break the rules.
This is where we need to be careful and help people to understand that, like my experience as a new poker player (and a new writer), learning those rules are important to learning how to create good data visualizations. It is only after you’ve learned the rules, understand them, and have gained enough experience and proficiency with the craft that you will know when it’s okay to break them.
The Role of Aesthetics
Finally, I want to discuss the role of aesthetics in data visualization as it relates to breaking the rules. In a 2018 Quartz article called The pie chart: Why data visualization’s greatest villain will never die, Dan Kopf wrote “The truth is that charts aren’t only about communicating data effectively. They’re also about aesthetics and drawing in an audience.” While the worst type of chart is one that misleads, another very bad chart is one that no one ever looks at.
If we are driven by a goal of getting attention, we may very well end up overemphasizing beauty and, in so doing, fail in our mission to accurately present the data.
While science often backs up our best practices, other science also points to the importance of our audience. For example, as mentioned by Mr. Kopf, studies have shown that people are innately attracted to circles — that’s part of the reason many people still like pie charts. Thus, there are times when we may choose to stray away from a best practice in order to deliver the aesthetics needed to draw in our audience.
For example, let’s discuss the dreaded gauge chart.
Pretty much all of us agree that it’s not the most effective way to show information. Other charts, such as bullet charts, are better in almost every way. So, why is it that executives still want gauge charts? It could be that these executives simply aren’t data literate enough to know that they’re bad. But it could also be that gauges are more visually pleasing and because executives have an existing mental model that allows them to automatically understand them (if you’ve driven a car, you know how a gauge works). Sometimes, we just need to get our audience to the table, and if that means breaking a best practice or two temporarily, then it may be worth it in the long run. Over time, as our audience becomes more comfortable with data visualization techniques, we can start to guide them towards less-familiar visuals.
Of course, we need to be careful here. There is a fine line between doing something to draw in our audience and doing something simply to get noticed. If we are driven by a goal of getting attention, we may very well end up overemphasizing beauty and, in so doing, fail in our mission to accurately present the data. As with everything, we need to carefully consider our audience, use case, and the overarching goals of the project.
Best practices are the guideposts that prevent us from going too far astray, but they should not be seen as hard and fast rules that should be followed at all costs. There are times when it is okay to break the rules. But it’s important for us to first learn and master them first so that, when we do break them, we do so in a thoughtful and purposeful way.
Thanks for reading!