Join Bill Shander for an in-depth discussion in this video Sketching and wireframing, part of Designing a Data Visualization.
- Now that we understand our data, and we know what story we're trying to tell, it's time to start sketching out some ideas. And this is a really critical moment in the process. You don't want to underestimate the importance of drawing. You want to spend more time doing this in fact, than you think you need to. Everything you can solve on pen and paper, will be a problem that you don't have to deal with in code, or in Photoshop, or in Illustrator. Now, some of you may be reluctant to draw, and I get it. Those of you who are not, you know, full-time designers, or you're not illustrators, you don't think you can draw, etc.
The fact of the matter is, that at some point in your life, you thought you were a great artist. Every short drawing you showed your mother when you were four, or five years old, they treated you like you were Michaelangelo. But then you grew up, and you had a brat sitting next to you in first grade, and that person convinced you that you couldn't draw. And maybe you're not gonna make a living as a fine artist, but you can draw, okay? You can capture ideas in visual form. Everyone can, and you have to do this when doing data visualization, so you can experiment and try new things, and fail quickly before finding the best solutions.
You can hear me rant more about why sketching is so important, and a bit about how to do it, in my Data Visualization Fundamentals Course. But here, I really want to show you some of the sketches that I came up with for this Human Development Index Visualization project. And as you'll see, I really can't draw to save my life either, okay? My lines aren't straight, my handwriting is terrible. I can't draw a circle to save my life. But I can totally conceptualize ideas visually, and again that's what counts here. So first, the big idea. I knew I was going to create a large illustration of an entire dataset, and that would end up as this vertical list of 187 countries.
OK, I knew I had several primary things to visualize. The overall HDI score, those three components that make up the score, as well as trends over time and inequality. And so I started sketching out these various approaches to how to get all this data on the screen. So, the main idea here, this was I think one of the first drawings that I came up with. Was maybe, I would have one big dot. Maybe the green dot, on those left-hand sketches, that was going to be the overall score.
And then these three differently colored dots, maybe somehow would represent the individual components. Maybe, the third one down on the left, would be some way of showing inequality. Right, there's like, the green dot to the right that's somehow showing that inequality faded back to the left or something. And so I was just sort of playing with ideas here. On the top right-hand one, I was thinking about maybe I would connect the dots with lines so you could see this trendline over time of how the components change from one country to the next.
So you can see, I'm just really playing with ideas here. And none of these worked, they're all kind of terrible. But they were the beginnings of something. This is a failure-free space to just try things. So the next set of drawings that I did, actually got me just about as close to the final result as any of the further drawings down the stretch. So, these are again variations on the idea. So in the upper left-hand corner, again, you can see I have some idea of the main HDI. Maybe inequality is the extension of those dots. And then, in that version, I thought that I might have the three components as sort of these stacked bars that lead to the full HDI.
But, that actually didn't make sense from a data standpoint. They don't really stack on top of each other like that. But on the right-hand side, I had what eventually became pretty close to the final concept. Right, I have three lines. Each one indicating the three components. All starting at the same left-hand point, and stretching to the right. Some bar that crosses them that indicates the overall HDI value. I was still trying to play with how to include inequality within those bars. But then I pretty quickly came to those split bars. If you could look to the left there, where you have those vertical black lines and then the little green bar going to the right of one, and then the red bar going to the left of the other.
Somehow putting inequality next to the three components in the HDI. And then to the left of that, there's little mini line-charts to indicate the trends. And as you can see, I was playing with different ideas there. So, maybe there would just be one trendline to show the overall change over time. Or maybe I had all three components as three different trendlines. But that was looking pretty messy, pretty fast. So, even as I was drawing this, I was getting the sense that this felt pretty good, and I thought this might be close to a good idea. But, you should never accept your first idea. Usually it's not necessarily in the best direction.
So I've started playing with other ideas, even though I felt happy with that one. So one idea, should I use sparklines? These little line charts to show changes over time. Maybe I include all four. The overall score, plus the component scores, but spread them out, so they're next to each other. Is there something interesting to look at there? And, I didn't think there was, so I moved on. I tried another approach to the three components and the HDI. Should the, maybe the three components shouldn't be bars. They should be just a different visual approach. Maybe they should be circles, and where the size indicates something.
And then have the HDI to the right. Maybe have inequality built into that measure over there. And again, I wasn't sure if this was the right thing, but I'm just playing around. I tried different shapes and approaches, so I had circular shapes and if you saw my Data Visualization Fundamentals course, you know how dangerous circles can be, but also how seductive they can be. And I was seduced enough to want to try a circular shape, but smart enough to know that this wasn't working very well. So, I moved on pretty quickly from here. Circles are very hard for humans to parse, and compare things.
So, dangerous thing to try, but worth trying. So I did more experimentation with shapes. Could I use triangles or diamonds? Or circles in some different way? Could I use cube shapes? Could the overall area of these things be understood visually to an audience? And so, I'm just playing with different ideas. And eventually, I came to this one, which I really liked the idea of. Again, I have my three components. They're kind of stacked next to each other. But again, I didn't think this would be easy to parse. It would be hard to see how a country at the top of the list, how their life expectancies, say that red box, compared to another one way at the bottom of the list.
I thought there just might be too much visual clutter to get away with this across the entire design. And so again, not being satisfied to take what I still thought was my best idea, I was trying all kinds of different things here. Alright, some sort of dot plot, where all the different scores are dotted across different parts of the screen. And again, I thought it made it really confusing, and very difficult to see how the different components in this case, would compare to the overall score in black. So, I kept coming back to that original idea. And pretty quickly, I realized yes, I think I'm gonna end up with a whole bunch of horizontal lines.
Those are those black lines on the left, which have stuff going on in them. They're gonna go all the way from the top of this thing, all the way down to the bottom. And then, what this sketch is, is just the beginnings of an idea of how to deal with the regional data. I knew that in addition to showing all of the countries' data, I did want to factor in some regional information. So I knew I was gonna have these little boxes, which might have, maybe a little picture of the continent. And maybe little line charts to show the trends of the region over time. Maybe some indication of inequality in the regions, etc.
Maybe, there would be lines from the region boxes connecting to the country boxes. But as you'll see in the final design, I abandoned almost all that because it was too visually crazy. Really, the sketching process was very helpful for me. I consider it very rapid, and as I said, risk-free experimentation opportunity, where you can really play with things. You can play with shapes, and how to connect things, and how to use space. And how to use standard or non-standard forms to get across the data that you wanna get across. As I said, my initial ideas were in some ways the best, which sometimes happens.
In fact, a lot of times you're tempted to go a little bit off the reservation, with weirder ideas. And sometimes, the weirder ideas just don't work. And sometimes they are the best thing you come up with. But in this case, my more standard, more simpler ideas, really seemed to work the best. So by the end of the sketching, a couple of hours over a couple of different days, I came to an idea that I was pretty sure was gonna be the solid concept. And once I had that idea, I felt comfortable, even though it wasn't fully flushed out on the whiteboard. I felt comfortable moving on to Illustrator to start really designing this thing, really getting the details down.
The UN Human Development Index used in this course is a perfect example: a composite number used to rank countries on how well they're doing across a range of measures (such as health, wealth, and education). Instructor Bill Shander shows how to make this index data tangible and approachable by imagining the story and visual approach first. He then builds the design in Adobe Illustrator, capitalizing on some automation and scripting abilities the program offers. Start watching for unique insights into the entire data visualization process.
- Working with the data
- Sketching and wireframing your design
- Roughing out the visual design components
- Manually creating the design in Illustrator
- Using Illustrator scripting to improve accuracy, speed, and repeatability
- Designing callout boxes, legends, labels, and more