In January, I was inspired by the amount of generative art being produced, in particular on my twitter feed. The hashtag #genuary2021 was full of examples of generative art, with each day having a different “rule” to inspire their participants. It’s hard to pick a favourite given there were 31 days’ worth of contributions from many different participants , but here is an example of a submission I really enjoyed, from Thomas Lin Pedersen (the rule for that particular day was “500 lines”.
Throughout the month I wondered whether I could recreate generative art using my tool of choice, Tableau. To further investigate the idea I needed to understand a bit more what generative art is. Wikipedia defines it as “art that in whole or in part has been created with the use of an autonomous system” – the implication being that it is created by something non-human, which in itself determines features of the artwork without human intervention. In other words, in the context of generative art shown on twitter, we are talking computer generated and random/algorithm driven (excuse my over-simplification and apologies if the assumption isn’t technically valid, that’s essentially the way I was viewing it).
I’d been inspired by generative art before – the work of Saskia Freeke inspired my Hitchhiker’s Guide visualisation below (as detailed in my previous post) but however abstract-looking (or random-looking) my data art style visualisations have been, I’ve always preferred to stay true to my data visualisation roots and generate output using data, however trivial.
With these ideas in mind, two friends independently started conversations with me – they had discovered Truchet tiles and instantly knew that they had found something right up my street! Adam Green shares my enthusiasm for eye-catching geometric disaggregated vizzes, and has his fantastic Alphabet Project right here. And Michelle Gaudette who collaborated with me on one my favourite pieces right here inspires me with ideas all the time knowing my design preferences and tastes in visualisation. These people both saw Truchet tiles and thought of me and the possibilities for future visualisations. But what are Truchet tiles and how could I use them for my next data visualisation idea?
Truchet tiles are tiles which are not fully rotationally symmetric that are designed so that when placed in squarely in a grid they generate interconnecting patterns. If the tiles are placed systematically then the new design may look equally regular, but when placed seemingly randomly then new random patterns emerge.
The most famous and simple Truchet tiles are known as Smith tiles – two diametrically opposed quarter-circles as shown below:
So by this point I have decided that Truchet tiles are going to be my entry into a generative art style visualisation. I don’t think it qualifies as generative art per se, since it will still be designed by me and generated by data, but my graphic generated from Truchet tiles will definitely be inspired by it. And ideally for me (and for anyone who wants to do something similar), I can create these tiles in Tableau in exactly the same way that I love to create visualisations. By creating a grid of squares small multiple style and by designing, creating and positioning my own polygons, I can actually create the visualisation in much the same way I create my flower visualisations. This is something I was delighted to talk about when speaking recently for Tableau’s #VizConnect series. Check out how I expand on the blog post in my talk below, including my design ideas and a Tableau “how to”.
I know now that I have the idea on how to go about this technically – but I wanted to look for something with more variety than the Smith tiles, since Smith tiles can be placed in two different orientations only. When I came across the following page from Christopher Carlson looking into more complex examples, then. I knew I had my source of inspiration: https://christophercarlson.com/portfolio/multi-scale-truchet-patterns/.
Carlson designs his tiles in order to use multiple sizes in multi-scale Truchet patterns. That was my first idea to replicate in Tableau, but its complexity means that might be a task for another day! But even within the constraints of using tiles the same shape as each other, he designs 15 separate (overlapping) tiles which tile together in grid form, which allows for much more variety in resulting pattern than the two simple Smith tiles.
I mentioned above that my visualisation would remain data-driven, however random or generative it would appear in nature. So my final question was around the underlying data – what could generate its own Truchet tile grid? And then could we see similarities, differences and patterns in tile grids generated using different data from the same category? I decided on opening passages from famous books, poems and songs. I could encode each of the first sixty-four words into a number 1 to 15 to select a Truchet tile. And, in true almost-random style, I would use the Scrabble score of each word to choose the tile. A pseudo-random generated tile for each work of literature. Inspired by Shakespeare, Orwell, Seuss and the Beatles (among others), I then had the data I needed.
Below is my final version – starting with the “how to read” page in order to explain the fifteen different Truchet tiles designed by Carlson and explained above.
And below, an example of one of my generated tiles (in both black on white and white on black, since I couldn’t decide which I liked best!) – this uses the first 64 words of Lewis Carroll’s Jabberwocky.
Click through in order to use the dropdown menu on Tableau Public and see the differences when different literary sources are chosen. Can you see patterns emerging in Dr. Seuss’s finest literary tomes? (Spoiler alert … a little bit …!)
I don’t plan to go down the full generative art route any time soon – I’m not sure I have the technical skills, but I will always be both interested in generative art and inspired by it as a design idea for data art style visualisations.