A couple weeks ago, a company called Hive published a ranking of the top 500 dataviz twitter accounts. They created the list using their own algorithm called “Peoplerank”. My own account (@W_R_Chase, please follow me 😘) was way down at number 366. I thought this whole thing was a rather elaborate way for Hive to tell me my tweets suck, I mean people tell me that all the time, and they’ve never needed a fancy algorithm to figure it out! So, clearly this list has no credibility since I wasn’t in the top 10, but it did reveal an interesting connection: several generative artists made the list. And I’m not just talking people like me who dabble in generative art, I mean heavy-hitters like Zach Lieberman, @inconvergent, @beesandbombs, Frederik Vanhoutte, and Jessica Rosenkrantz.
Perhaps this is some peculiarity of the algorithm, but I think it reveals a real connection. At a talk I gave last year on creative coding I showed what I call the dataviz <–> generative art spectrum.
I think about this spectrum primarily as the combination of two variables: the input and the intended outcome. On the far left we have pure dataviz, where the input is structured data that represents real-world observations, and the desired outcome is to clearly communicate those data and their meaning to the audience. As we move towards the center the input remains structured data, but the desired outcome shifts slightly. Our goal is often still to represent and communicate some information, but we have another goal of creating something beautiful and artistic, and this goal may begin to encroach upon the communication aspect. Perhaps now we choose a visual representation that isn’t technically as clear and accurate, but it’s more beautiful and engaging. As we move further towards the generative art side of the spectrum, our goal shifts more and more towards creating something beautiful, and away from a faithful representation of data. The input also shifts at this point, it may be somewhere in-between: data that has been altered with randomness. But as we get entirely to the generative side we do away with data as the input and shift to something randomly generated.
The reason I emphasize this spectrum so much is that it has a lot to do with communication design. There’s a lot of discussion in dataviz around which charts are “best” and that usually boils down to “simplest and most clear”, but this discussion leaves out factors like engagement, emotional connection, wow factor, and memorability. These are all critical factors for effective communication, and they’re precisely what you gain by moving a little bit away from the data side, and a bit more towards the art end.
I’m not advocating that you use a generative layout for your KPI dashboard, but if you’re preparing a shareholder report, a little wow factor and memorability from a more artistic graphic might be just what you need for that lay-person audience. As with any design decision, there’s a tradeoff between the dataviz and generative art side of the spectrum, you sacrifice either clarity or wow factor; simplicity or memorability; accuracy or engagement. That’s not to say you can’t have a beautiful and accurate graphic, but there will always be decisions you could make that will shift your design one way or the other.
Deciding where your graphic falls on that spectrum is entirely about audience. If you’re designing an internal government report, clarity and accuracy are the priority, so go for the bar chart (or just use a table). But if you’re reporting the same statistics in a newspaper story, your job isn’t just to convey the information, it’s to frame it and tell a story that readers will find engaging (so they actually finish reading it) and memorable (so they tell their friends about it and spread your message). So next time you sit down to create a visualization, don’t ask “What is the best chart?”, instead ask “Who is my audience, and how can I design this to communicate most effectively with them?”
I get a lot of my inspiration from generative art, so I wanted to pass some of that along. My goal with this list is to highlight people that might be up-and-coming, or represent a more diverse segment of the generative art community–people you may not have heard of, but they’re making some of my favorite work out there.
Sofia Crespo makes speculative biological creatures using neural networks, just check out this sample from her incredible Unnatural History series
Helena Sarin also makes generative art using neural networks, often combining traditional art with the natural world and feeding it into a computer, her art is surreal and beautiful
Dmitri Cherniak is making some of the most beautiful art on Twitter today. Incredible textures and always fresh ideas
Deaconbatch is a member of the thriving Japanese generative art community. Constantly experimenting, constantly sharing, always inspiring. I particularly love this recent series:
Saskia Freeke is an absolute creative coding machine. For five years (five!!) she has produced daily artwork. I’ve tried daily challenges before and never made it more than like a week, so I’m incredibly impressed she’s made it nearly 2,000 days. Also her art is beautiful and it’s wonderful to watch how systems evolve
Way back in The Before Times, I started this project called 12 Months of aRt where each month I made a new generative art system in R and blogged about it. I actually managed to get most of them in on time, but the final month slipped away from me, and then COVID hit and my inspiration and will to focus just dried up. But I’m happy to say that now I’ve finally finished the last project, which you can see and read about on my blog. I like to think you as my subscribers get some special early access, but the truth is I don’t have time to figure out how to post this on a private platform only you can see, so it’s just out in the open, but at least you all get to hear about it first (I won’t be announcing this publicly for a couple days).
This month I developed a system based on grids that distort as they pass through polygons. The project ended up taking far more time than I anticipated, and in the end it was really only 90% finished, there’s still a few bugs and missing features. But as my schedule piled up, I decided it was just time to call it a day and ship it. I do enjoy the output, and I think it has a lot more potential with some small additions. The most exciting part was that I finally broke out my pen plotter for this project, and the plotter continues to be the most mesmerizing thing I own.
In my last newsletter I invited you all to respond and tell me a little about yourselves, and I was so touched by how many of you reached out. I had some really interesting conversations and I loved getting to know more about my readers. Several of you also sent in some interesting articles, I can’t share them all, but here’s a few that I loved
Evelina Judeikytė told me about a wonderful article she wrote on how public speaking principles can improve your designs and I totally agree, I loved it!
Corrie Bar was reading this wonderful essay from Wired on storytelling in visualizations
Juan Franceschini sent me this essay on meetings and the manager’s schedule vs. the maker’s schedule by Paul Graham. The piece is now more than a decade old but feels more timely and relevant than ever in our post-pandemic work-from-home world. As someone who has recently joined the business world, this essay really struck a tone with me. I find it really hard to get any difficult work done without significant blocks of time, but I’ve learned that remote work in the business world goes by the manager’s schedule, and I often find my days pierced with stand-ups, all-hands, knowledge-shares, and “quick calls”. These frequent interruptions really throw me off and prevent me from starting big tasks that require a lot of focus and energy, which is basically the whole point of Paul’s essay. I don’t know if I’ll be able to make any change in my own case, but at least now I have the perfect encapsulation of my problem: I run on the Maker’s schedule, my workplace runs on the Manager’s schedule.