Skill Level Appropriate for all
- Data visualization is breaking the bounds of the printed page and the screen in many different ways. People are making data visualization objects using 3D printers and even Play-Doh. They're mapping their runs and using that data to spell out marriage proposals. Swedish academic Hans Rosling famously used objects like income inequality and population growth. The tangibility of real-world objects is very powerful. Humans can relate to objects we can touch and hold in a way that we can't quite relate to in the subtle abstractions of pencil marks on paper or pixels on a screen.
Humans value physical objects for a host of reasons. For one, there's a difference in how we perceive things that we own as opposed to things that we may have access to. So a physical book sits on my shelf. I truly own it, whereas an E-book is something I can use, but it doesn't really feel like it's mine. It feels like I'm just sort of renting it. I deal a lot with creating visualizations in a virtual space. However, for this episode I want to focus on the other end of that same spectrum, which is respecting and working with the physical dimension to create data visualizations, even when the final product ends up in print or on screen.
This is all about using your hands when brainstorming and thinking about the physical world when it's relevant to the data story you're telling. I always tell my workshop students to close their computers and pick up pencil and paper or a whiteboard when coming up with ideas for visualizations. In fact, when I do in-person workshops, I bring personal mini nine by 12 whiteboards for every attendee, which they get to keep, because I really believe very strongly in this process. Here's the thing.
If you're waiting for your computer to catch up to your thoughts or you're fighting software features, you're going to struggle to move quickly and experiment. With a whiteboard or an iPad with simple sketching software, you can come up with a concept and you just immediately sketch out your idea, and you can almost instantaneously think about whether it'll work. This physical experience is part of what I'm talking about. Also there's the idea of embodying physicality in your data. So I just finished a visualization about computer keyboard layouts, okay? You know how your keyboard probably uses the QWERTY layout? There are actually a lot of alternatives, okay? The question is why do we use QWERTY? Is it the best layout for speed and accuracy? Should we consider alternatives? What alternatives might work better? And how do we even measure these things aside from testing how people perform typing on different layouts? One of the ideas was about how you need to jump from one key to another while typing.
So, for instance, if you type the word any, the n and the y require you to jump two rows on the QWERTY keyboard with the same finger. So letter combinations and the amount of calisthenics required to achieve them are probably worth measuring since it seems logical that if we could identify common letter combinations and minimize movement, it might lead to smarter keyboard layout designs. So I was thinking about how to measure that, and I came up with the idea that I would write code to essentially read an entire book of text and the code would analyze the text and then look at the most common two-letter combinations.
And obviously I'm trying to show a visual representation of the movements required to make those combos happen. So on my whiteboard or my iPad I start sketching. I'm just trying to come up with a good idea. And so, you know, what am I going to do? I'm thinking to myself, all right, basic simple idea. Maybe I literally just have all the letters in order. And why it's alphabetical, I don't know. But I'm just sketching ideas here. And maybe each time there's a jump from, like, one row to another row, I'm just indicating it with dots or lines or whatever.
Some letters have a lot of jumps. Some letters have very few. This is data visualization. But instead of actually doing letters, why not make them look like keyboard keys? That's one part of the idea. And then maybe what I want to do is actually show the jumps a little bit more graphically. So I can have, like, little connecting lines. So this key jumps to this key. This key jumps to this key, but this one jumps a lot more so it's a nice thick line. I could even use both sides of the index here.
On one side it's keys that are jumping from the top row down. On the other side maybe it's keys that are jumping up. And I could use color, you know, red for down, green for up, who knows? This is an idea, but this is a keyboard. So why am I drawing these in alphabetical order, in this case sort of vertical order? I don't know; it doesn't really necessarily make sense. So I immediately start thinking about, well, what if I actually made a keyboard layout? And, you know, sketching quickly, I don't want to waste too much time doing that, so I'm just going to do circles. If it looks like a keyboard, we're talking about jumping from key to key on a keyboard, it would make sense, and it might look like this.
And now I can do my connecting lines like this. And now when you're connecting between these two letters, people are just going to get it. They're going to understand what it means. It just seems like a logical way to go. Now, these three ideas, and of course, there are dozens more, as I said earlier, a couple of them are sort of data visualizationy, but they're not about physicality. Now, that last idea, this is sort of physical, right? It's the physicality of the keyboard. But I'm talking about something a little bit bigger. And so I'm trying to get at, you know, really expressing physicality and allowing people to have a very tangible understanding of the issue that I'm talking about.
So how about the idea that these jumps represent real movement that adds up to an actual cumulative effect? So when I was doing this project, I was using Anna Karenina, okay? That book has over two million characters, okay? Literally, if you type out the whole book, it's more than two million characters. Now, on a QWERTY keyboard, these jumps that I'm talking about from the bottom row to the top row or back, there's over 140,000 of those jumps. Each jump is about an inch and a half.
That works out to over three miles of travel an inch and a half at a time, okay? So that adds up to a lot of extra typing. So I could try an idea like this. I'm thinking to myself how can I express that idea in a way that'll make sense to people? So I'm realizing, well, you know, 3 1/2 miles, that's a long ways. What's 3 1/2 miles or something sort of relevant that people could relate to? And so I'm thinking, well, you know what? If you go to Washington, D.C., and you were to walk from the Lincoln Memorial, okay, this is Abe Lincoln right here, and if you were to walk from there to the U.S. Capitol Building, okay, that's about 2.3 miles, something like that.
So literally I could do a visualization like this where I show those two things, and literally you're walking and we show footsteps or something like that, walking all the way there and walking halfway back. So this sketch, this idea, very quick and easy to sort of generate this concept. I'm able to iterate a bunch of different ideas across, you know, all four of these ideas that I came up with, all of which have physical dimensions but none really better than walking the National Mall. Now, you could argue that only those who have been to D.C. would get the reference, but I could make other references to distance to make that point.
So work with your hands. Work quickly and iterate and embrace the physical world when thinking about data that has or relates to physical dimensions in any way. Up next I'll talk to someone who's doing visualizations using real-world objects. Her name is Michelle Reale, and you can check out her work on Instagram. Let's get to it.