▲ 182 points
back
61 comments
During the rise of the generative AI backlash, I once saw a take along the lines of "data visualization doesn't count as art because you write it with code: art only counts if you put sweat and tears and labor into it."
I was too baffled to respond.
I bet this person has never created a data visualization.
On the contrary, I think data visualization (along with architecture, mathematics, and mechanical watches) is one of the most interesting art forms that requires strict constraints.
By that logic, photography, digital painting, music production, and even architectural design wouldn't count as art because they involve tools and technology
> And creating art is about more than just precision. It's about evoking emotion, telling stories, and sparking curiosity. It's about making people think and feel.
For sure, the best artists are the ones who take their whole perception of stuff, with the feelings and everything else, and put it directly in someone else's brain. I would maybe say that some data visualization is supposed to do that, particularly data journalism, but definitely not all data visualization.
I think not all data visualization is art in the same sense that not all writing is literature, and in fact almost all isn't, because it's more like "I am in a shared info network with my boss/colleagues/clients/readers, and we need to communicate as rapidly and directly as possible, so we've agreed on a shared visual language that is higher bandwidth." Emails are slightly less exciting to read than a novel but like a million times faster to write per word.
Which is kind of sad. It would be fun to hone that skill and there is no room for that at my job.
(Side note, it seems like everyone always says that the pie chart problem is about lengths vs angles... but isn't it just that one chart has a shared baseline and the other doesn't? Taking the middle out [donut chart] to emphasize arc length doesn't actually solve the problem. Conversely, a stacked bar is more like a pie chart than a bar chart, except with a weird aspect ratio if you only have one stacked bar).
Author of the article here. Great points, thanks for reading and commenting.
I think that data journalism is a great example of that "put it in someone's brain" idea.
I also think there are all kinds of places we're seeing more novel approaches in too: data art, physical data viz (e.g. museum exhibits), visualizing huge datasets, etc. The Climate Stripes project, maybe the most famous data viz of the last couple decades, was made by a climate scientist.
But I definitely agree that this "more experimental than rigid" side of data viz doesn't capture the majority of the field, nor should it!
I just wrote a book (released yesterday!), Effective Visualization, that takes almost the opposite approach. Follow well-known patterns and know your audience. That should guide you to telling a good story.
It can look pretty, but making it art for the sake of art probably won't resonate with your audience.
That's been my experience after I've trained and consulted with some of the biggest companies in the world.
Congratulations on the book drop!
So I'm not sure if you read my piece or just wanted to drop a promotion, but I think you're misrepresenting my view here. I'm not advocating for "art for the sake of art" and I certainly don't think you should not know your audience.
My main point is that I don't think all of data viz is as simple as you're implying here, i.e. that a pretty chart probably won't "resonate" or that a chart from the Tufte school of thought automatically will.
Almost all pictures on that page are cropped weirdly when viewing on mobile. Words are cut off in many.
No, it's not an Art.
I get that people use the word art as "difficult", "obscure" or "intangible" but...
Art is about self-expression, evoking emotions, and open interpretation. Design is about problem-solving, functionality, and clear communication.
Clearly DataViz is Design.
I don't see art and design as mutually exclusive. I also don't think that data viz is exclusively about functionality.
Take a project like this: https://www.dear-data.com/theproject... this is clearly data visualization and, to me, quite evocative. These visualizations aren't designed for clarity and they don't need to be, that wasn't the goal.
> Art is about self-expression, evoking emotions, and open interpretation
Art is about these things, but is also about many other things. And art can be clear in its intentions, leaving little to interpret.
What is the purpose of data visualization? Often to evoke emotions and to help someone not familiar with the data understand how to interpret it.
The people who accomplish this most effectively understand that the point isn't just to force data points into a visual form. If it were so simple, more people would be good at it (they aren't).
> Clearly DataViz is Design
I don't understand this sentence. I'm not trying to be difficult, but if DataViz is Design, what is Design?
Art as in craft/field! "State of the Art" is a phrase that refers to what is possible within a domain. There is also the implication that it takes judgement, maybe even an artistic sensibility. This is in opposition to a strain of thought within data viz from the Business Intelligence/Analytics side of the field that has succumbed to the MBA/Process gospel that preaches Repeatable Process and the siren song of good data visualization without using your brain!
Is Design an Art?
there seems to be two competing visions of what is art:
art as a the top degree in any discipline; like in "state of the art"
but there's also art as that little part of science that escapes rationality, the self-expression drive of a human and so on
> "It can be a science, an art, and a craft all at once."
suggests the article leans for the art as the top level of some discipline definition of art
It's "art" as in the phrase "More of an art than science"
data Viz for it's own sake might be design - in real applications, you always have a story you're trying to tell - and it doesn't go away just because you're not consciously aware of it.
Ignoring the actual argument to focus on the object level example of pie-charts:
Pie Charts are good for showing relative proportions of a whole for a relatively small number of items. Donut charts are better because humans tend to misread area slightly in pie charts.
Author of the article here! Yes you're completely right. I think they also have other underrated properties, like they fit neatly in a square which can be helpful for laying out a dashboard. And for some reason, in my experience, people just like them.
But when done poorly, they can be a mess. Like any other chart.
I don’t love the “me against the world” framing, but maybe I’ve just been blessed with good teachers — this is a great summation of what the HCI faculty at GaTech espouse! If it’s not a consensus yet, it should be.
If the author pops by, I’ve just gotta say: you’re killin it with the cute illustrations, especially the dials one. Very inspiring stuff!
Thanks for sharing, OP.
EDIT: if I had to pick out any one idea for a casual reader to learn about, I’d definitely highlight “Cognitive Load”. It’s used as part of a broader discussion here, but when looked at in a certain light it can capture a whole lot of diverse factors in one consistent framework!
Thank you! Yeah, I hear you on the framing. I don't have a formal education on data viz myself (I guess not many people do?), but I've gotten into the world over the last few years. A lot of the literature I've seen has come across as dogmatic to me, which is what sparked this article. That's great to hear that your experience is more aligned with this way of thinking.
And thanks for the nice words about the animations! Glad to hear that work doesn't go unappreciated haha
I love the framing of data visualization as a balance of trade-offs. It's not just about picking the "best" chart type, it's about knowing why you're making the choices you are
On my old iPad 4, this page is essentially blank.
I try to apply principles of progressive enhancement when I want to communicate. I start with plain HTML, basic text block structure and images.
Interactive data viz can be a sublime way to tell a story with data. But if I can't explain the data with the basics, I try to rethink my approach.
I absolutely believe that data visualization, as an art, A craft, can transform the way we share essential ideas.
But I try to make sure a static hero image is up there when necessary.
I've grown out of being impressed by flashy and difficult to understand data visualizations. I'm much more impressed by novel insight, clearly presented, in familiar ways.
I think the article addresses your point: "The fact that some disciplines of an art form demand more precision doesn't mean the whole field shares those demands. It would be like saying best practices for photorealism should be applied to abstract expressionism, or how we approach technical writing should be the same as how we approach poetry."
It wouldn't make sense to say "I've grown out of being impressed by complex prose" - you're welcome to enjoy Faulker or not, but he simply didn't have the same goals as the authors of the Stripe documentation, and judging both pieces of writing by the same standard is basically pointless (except, perhaps, as an art project of its own).
I'd even apply this to coding itself. Day to day, I think everybody around me should be writing Blub, grug-brain code, but I'm happy there are people trying creative, weird languages and mind-bending ideas elsewhere and I'm curious what they find out.
I think publications practicing actual journalism like The Upshot and ProPublica do an incredible job of this. The focus is always clear presentation and accessibility above all else but its undeniable that so much of their work is also just gorgeous
I'd call what you describe as impressive, art.
Flashy trashy is just crap. Some might call it art but then again everything is art these days.
I think that's the point of the article. Novel insight is almost a pre-requisite to determining how to present the data in a way that's quickly and easily understandable.
this makes me nostalgic for the days when d3.js and observable made it regularly to HN front page...
Was looking for some mention of Mike Bostock and his epic odyssey into this space.
For those who aren't familiar https://observablehq.com/@mbostock
Be the change you wish to see in the world!
I actually got the pie chart almost 100% correct…
> It is completely reasonable to give up some precision to:
> - Reduce unnecessary cognitive load and emphasize a feeling, as in the Climate Stripes example.
> - Encode a much higher volume of data, as in the Climate Globe example.
> -Add delight and use a visual metaphor to show trends, as in the Coffee Shop example
This the most absurd claim I have ever seen regarding data visualization. Give up precision for delight? Give up precision for a feeling? This article should have been tile "How to lie with graphs". The article is advocating for how to use graphs in the service of rhetoric, not science. This is exactly why visualizing data should NOT be treated as art. The first and third bullet point are obsolutely not reasons to give up
Giving up precision is different to giving up the narrative. The narrative can be helped by engagement and if that comes at the cost of precision, then that was the cost of getting the message further than it may have otherwise reached.
In this case, what is meant by "give up some precision" is to try removing some sharp edges that get in the way of telling the story. It's a long way from lying (when done with due consideration).
They're examples. Choose what works for your data and you your story to get the best result.
You think the Climate Stripes are an example of "lying"?
99% of the time it is design, 1% it is art. The goal is always to convey information in a digestible way without being misleading. An example of it being art is that video that illustrates death counts in WWII.
The Wall Street Journal Guide to Information Graphics should be a required for reading for anyone working with numbers.
tldr: only use color when necessary, avoid bold typefaces, keep scales consistent across charts unless there is a specific need to do otherwise.
A few thoughts there:
1. I generally agree with you on that goal, but there are a TON of ways to convey information in a digestible way without being misleading. Even if you follow good practices.
2. "Convey" is doing a lot of heavy lifting there. What are you conveying? One number? Many numbers? A specific story? As much data as possible to let the viewer find a trend themselves? Are the numbers important or the general direction of the data? Picking an answer and making the visual serve that goal is an art IMO.
It’s so easy to measure how fast and accurate users are at completing real tasks in your domain using various viz. I’d be disappointed if the conversation was purely between rigid principles and artistry.
Yes. This is exactly the sort of information you’ll find in Bill Cleveland’s books.l (referenced at the start of the article then somewhat dismissed later). Bill Cleveland did all these sorts of studies with actual users about which visualizations are most accurate and quickest to parse. He did tons of work at IBM studying this stuff because they needed to use effective visualizations to convey information to decision makers. Unfortunately most of the visualization discourse has been taken over by people like Tufte who focus on style rather than measurable substance.
Visualization seems to have stagnated in the AI craze era. Lots of androids touching holographic screens and what not, as metaphors for the almighty "AGI" coming any day now, but cant recall a major new development in the last five years?
One area that might be "pregnant" for some new approaches is the visualization of large datasets, eg large graphs. Extracting useful (and objective) information instead of ovewhelming with the sheer number of data points. That is indeed the art of visualisation.
What was the major new development in visualization in the past 10 years?
I don’t think there has really been one in decades. We’ve gotten better tools for making viz interactive but that’s about it.
> The reason is that humans are much better at comparing lengths than they are at comparing angles.
Maybe the subdivisions with exact numbers also play a significant role in the effectiveness of the bar chart in the article. It lets you compare against a clear reference, so you're often comparing to the scale itself rather than directly between bars.
> The reason is that humans are much better at comparing lengths than they are at comparing angles.
Are you sure 10% increment guides in bar chart have nothing to do with this?
Hah, so I actually went back and forth on that. The original studies around visual inference also included axis labels on the bar chart. The fact that you CAN label the axis on a bar chart that way is actually an argument people make not to use pie charts.
6% error on that pie chart. Pretty proud of myself ;)
2% goml
Damn the coffe shop tracking ha
Nice. But please educate yourself on a difference between art and design.
They have similarities in language, but have completely opposite function.
Art is a personal expression with a given medium which translates directly the intent. Design has a focus on thinking about the function and perceptions of other people.
Visualizing data is a design problem.
No reason those need to be mutually exclusive. Posted this in another thread, but I don't see a good argument against this project (https://www.dear-data.com) being both data visualization and being an act of personal expression.