This interview of Andy Kirk contains a discussion of his ongoing blog posts about what he calls "The Little of Visualization Design". He shares some of his most interesting and inspiring examples of the small details that make huge differences in this work.
(upbeat music) - So I'm very happy today to have Andy Kirk with me. Andy is an independent data visualization designer specialist based in the UK. He has written a fantastic book, a handbook essentially, on data visualization. He conducts training sessions and he has a really great website, visualizingdata.com, which covers all kinds of data visualization topics. So, Andy, thank you very much for joining me here today to talk about the little details in data visualization design. - My pleasure, Bill. Great to be talking to you. - So, I don't want to waste too much time on this. It's not worth it. This little short interview, but we do have to have a starter for the very slight argument about the spelling of data visualization. It's a classic argument, I know you feel very strongly about it see so for those of you who don't know data visualization is spelled with a Z, visualize, visualization and in the UK, they spell it with an S, and we don't have to linger on it, but on the other hand, this is about details this interview, and so maybe it's worth talking briefly about that small detail, the S versus the Z. So Andy, what do you think, should we argue about this for a little while? - [Andy] Yeah I mean, we've only got a certain answer as you say Bill to argue over this. I mean, obviously it's a spelling choice that is very precious to our British sensibility. It's also the only available URL at the time be buying my URL Z with it so it was a natural choice to fall back on the S. But I guess it's the one area that I'm perhaps most relaxed about Bill in terms of labels and definitions of what we call this endeavor is it visualization does someone call their piece of work an infographic, a diagram an interactive. It's actually the most relaxed I am about the details of visualization. I don't care what you call your work so long as it's still motivated by a desire to help somebody understand something. So I'm not relaxed about this compared to many other things in which I am not relaxed. - (laughs) Alright, let's talk about some of those things. So you have a series of posts on your blog called the little of visualization design or the little of data visualization. I can't remember exactly how it's phrased. But the gist of it is It's all these small little details that people can change to make their work better, small little details that might trip you up, and I just want to start off by asking you, what was your motivation in creating that series of posts? - Basically a question, I think there's two parts to the answer. First of all, I think the small details are the things that make the biggest difference in visualization design because we are responsible as creators of these works. We are responsible for every single pixel. Every pixel matters. And guessing in contrast to some other creative endeavors like graphic design. Every pixel in visualization could potentially inform. Whether it's the color, the size the placement, the sequencing, the visual hierarchy, every attribute could potentially represent data or meaning as distinct a graphic design. For example for which there's more scope, let's say arbitrary choices. And that's not to put one down versus the other but they have just different pursuits. And so I do think you can create such wonderful visualization work by taking care of these small details. Every single element has to have some rationale, some justification behind it. So the first part of the reason for why I did this pursue, this series of post was because there are so many different ways we can look at this particular topic but also from a very, I guess mechanical coping mechanism point of view as somebody who does run a blog site, trying to find new topics can be quite hard sometimes. I've been doing this for 10 years. I think I'm nearly up to around 900 posts. And so what I was looking for was a repeatable theme that would allow me to do quick blog posts and not something will be very onerous every time I came to the writing of a blog post ask. And so by find something like this where I could look at any visualization particularly find a subject for this posting. It gave me something that I could then, repeat repeat repeat and I think I'm up to now, as of today 68 posts. And to be honest I don't really see where the end point is. I think it's something that can go on and on forever. Even with the same observations made about different pieces. - I think it's a great point. So there's a big difference between data visualization and graphic design in that as you said every pixel can provide meaning. And so in fact, I have a scatter plot with a thousand dots on it. That dot might be the size of a pixel. That pixel literally might be data. So that brings up the next question, which is you've done several dozen of these 68 and I'm wondering if you're seeing types of things that are dominating the category. So, you know patterns, the types of the oil is all about your color seems to dominate the issues or labeling or context setting information axes. Are there any patterns you're detecting? Or is it that you just sort of implied if it's an endless thing maybe it's all over the map. - In a sense it's both built because I think part of it is yes the in theory any piece that I encounter could have scores of potential little vis observations. What I've actually reflects upon having just browsed for my full series is that one of the most repeated themes concerns composition layout choices positioning, sizing, placement of axis scales, and labels, and titles, and just the layout of charts. And I think in part of that plus 'cause that's where my radar is perhaps more tuned to. And I do feel that in my own visualization technique maturity, small decisions about placement and the physical use of space has perhaps crept in as a real focus of my own work in the last couple of years perhaps. And so perhaps that's where my radar is mostly tuned to see in these astute choices about the use of empty space the deployment of containers to house and separate different contents, the innovative way of sorting even bar charts. Just the innovative way of sorting that goes into a more sophisticated realm than just alphabetical or most to least. So I guess I'm particularly not seeing these choices, but as you said color is a repeatable topic. When I started this put these posts and series, I actually started off from more of a negative perspective perhaps criticizing small choices that had been made. And very quickly I then realized that that's not aspirational. So I switched the focus to be more on celebrating good choices that I see so that these collections of post become something that is a best practice collection. Rather than just here's another bad observation look how this person has done this terribly badly. You know, we all make mistakes. We all make bad choices even now after 12 years in this field I do. So, I've switched it and I think from that point of view, I think some of the more sophisticated uses of space and color become often the separator between good and great visualization. And I think that's perhaps where my lenses dominated over the last couple of years I have been in this Pulsing Series. - There are plenty of opportunities to find people criticizing others work. So I think focusing on positive is never a bad choice. - [Andy] Absolutely. - In that process, the especially thinking about inspirational examples. Have you come across any big Aha's? Any sort of like wow this one post that you did that revealed this the small detail that was like wow. Maybe something you apply every day in you work. Now you can't see it once you notice that small detail. - I think the, and it's not just one actually, but the several ahas. I've come from small details that involved almost this integration or merger of two traditionally distinct features. So for example, it might be the use of a color Legend within a title. So rather than having a title that includes mentions of the things that are colored in the chart and having a separate legend that also does that. The legend and the title are merged into one single thing. I've seen titles that are against the backdrop of a stacked bar chart because the title shows or describes how one thing dominates over something else. I think the example in my blog series was about how men dominate female in one particular topic. I don't remember the exact topic. And so the title revealed that finding but then behind the title was a stacked bar chart that kind of encoded that in a decorative background coloring. The use of a color Legend that also doubles up as a histogram to show the distribution within the map that it relates to of those specific discrete colorings. So it's these features that remove redundancy. They sure to incorporate two features into one. And I just think that's kind of quite advanced thinking whereby you can look at a piece of work through fresh lens and think Tony both of those things are can I find a way to merge them together? So yeah, there's lots of very talented people out there who are now deploying these things together and you know in my own work as you said, I'm trying to employ some of these tactics. - Yeah, that was one of my questions that I had for you. So you're discovering all these great things you are a data visualization practitioner. How often and are you implementing them? Which sounds like you are be are you constantly thinking about these small things and it's almost become a part of your checklist. Every time I'm doing a bar chart about apples and bananas that text is going to be in red and yellow or whatever it is. Or is it more like you're working on something you sort of notice this is an opportunity to do a better job. How's that work for you? Yes, it works in two perspectives. So first of all, yes in my own work, I do have this almost framework checklist to ensure that I've covered all bases. And this extends also to how I teach and instruct the subjects because the way that I try and dissect the anatomy of any visualization is across five distinct elements. You've got the representation layer, which affect even charts. You've got the potential for interactivity those features. You've got the annotation which is the element of assistance headings, titles abscess scales, and then you've got color choices and as I've just mentioned the composition choices. So whenever I'm thinking about my own work or teaching the subject I break these discussions down to those five discrete elements. And then what I've found is that there are three key principles that help me to determine whether across those five elements of design, I've made good choices. And those three things are is my work trustworthy? So do people believe the accuracy of what I'm portraying and the contents and how I've portrayed it. No distortions are no lies. Secondly, is it accessible? Am I giving people sufficiently easy access to the understanding? Which may not mean it's simple it may not mean it's necessarily quick, but I'm trying to limit the any unnecessary obstacles to understanding. So do my choices ensure there is efficiency, optimization in the readability of that work. And then thirdly is it elegant? I think this is the hardest one to articulate often because shot and notice elegance through its absence. We notice inelegant things ugly things chaotic non-unifying things. And I often find that third one is where a lot of these little vis choices land. Squeezing out those extra small details that create that sense of elegance whereby someone just takes it for granted as being aesthetically pleasing and sustaining that appeal throughout the process. And so these five distinct design elements against the lens of these three principles. Give me my own checklist for doing and for teaching. But what it does mean is it sometimes I lack discipline in finishing work. If it's a piece of work that does not have a nice defined time scale or deadline. I go through this endless process of iteration and iteration and I struggle to finish. I do struggle to stop finishing piece of work and that's why I often find most comfort when I do have constraints and boundaries and things I can't just endless to pursue every spectrum of options. So yeah, it does come into my work but sometimes in a good way and sometimes in a bad way just through the sheer inertial choice. - Don't let perfect be the enemy of good. You have to sort of have that written on above your computer somewhere. So, you just brought up you do teach data visualization. And so I'm hesitant to do this but I have to ask you, so this series that I do a lesson on the theme and then I have an interview on the theme. And so I did a lesson on this theme and I'm wondering if you were to teach a lesson on the little of visualization design those small details a short five-minute lesson. What would that be and you have to give the lesson right now on the spot but more like what do you think you would say in 20 seconds to say this is the secret to getting those details, right? - Often the basis of this observation is that to be a good writer you need to be a survey reader. To be a good creator of visualizations you need to be a good consummate reader of visualizations and critique of visualizations. And I think the biggest method that I used to try and impart this is to get people to look and dissect any visualization they encounter. And to just get them to ask questions about every single element that they can see. to first of all list, literally list all the contents, all the color choices that have been made across every single content. Every chart that was used every label. Every decision that you can see evidence of in terms of placement and then ask questions about are they effective choices? If they are fantastic, but why? And if they are not what would you do differently? 'Cause it's very easy to criticize other works, but you got to back that up with and instead I would have done X Y or Z. So I think just simply just reading work. You don't need to be even a skilled creator, but just look at any realization and dissect it across these five elements I mentioned there and ask questions about what works what doesn't work. And what would you do better? - Andy you sort of implied also intentionality. Intentionality about every decision that you make because you can throw stuff together. You can say Excel give me a chart and if you're not being intentional about those decisions, then a lot of those details are not going to be great. I also have to say that I have to step back because you mentioned a word earlier and as I was thinking about how I wanted to ask this question that occurred to me, I was going to phrase it this way, you know, there was this great conversation on the data visualization slack community about elegance and defining elegance. And then it occurred to me was that your, did you lead that conversation? - [Andy] I did. - So for those of you who don't know the data visualization Society is an organization, essentially like an association that has been created in the last year. And there's a great slack channel conversation. One of the channels is about topics just sort of focuses of discussion. And so yeah, you lead a great conversation about elegance. And it is so hard to define but I agree it is all about those little tiny details. And making them all as perfect as you can without driving yourself insane hopefully in the process. - Absolutely and I asked that conversation because I do think one of the things that we can learn in visualization is lots of the techniques that come from other worlds. Whether it's things like video game design, whether it's things like architecture even cartoon design. How you impart comedy and humor make a point in a single cell a static piece of work often with the very time prescience context. These are all things that we can learn from. And so elegance is something that I feel just permeating so different other pursuits. Whether it's movie design, set design, even just things like the elegance of comedy. The brevity of a beautifully structured joke is fantastic. So I wanted to have that conversation with other people in the slack community of the DVS because they come from everywhere. They come from every different discipline and background and domain area but also different parts the world. Elegance is to how we in northern Europe see things is different to how elegant aesthetics exist in the Far East and in South America and different regions of the world. So I wanted to use that conversation to bring together the common themes that structure this quite elusive and enigmatic pursuit. And hopefully there are some points in there which will give people a sense of guidance to just enhance the work they are producing. - That's great. We're running a little bit low on time here. These interviews are pretty short, but I was wondering if you had any final thoughts you'd like to share about getting those details just right in data visualization. - I think that the biggest advice to just add on what I mentioned before which is just expose yourself to as many different examples and techniques and works that are out there. Jump on Twitter, go through different blogs. Go through different books and you just see every type of work that's out there. Both the contemporary works, the recent classics from the in our last decade but also then going back to the 1900s, the 1800s and just expose yourself to as many different pieces possible. And especially if you are somebody who let's say is working in very much a corporate environment exposure to self to works that's not copy paste that is more artistic and creative. That may not be entirely relevant to your day job, of course, but they'll be themes about layout and labeling and color uses that you can draw inspiration from. So even if things look very different to how you might expect to be cheap, but using your work, you can still learn from that. Both positively and negatively. So yeah just open your eyes and eventually it'll become a curse. Because you'll stop being up to see these things as an everyday reader and everyday punter. Just trying to be informed you'll start to see these as design choices and that's something that hopefully will elevate everyone's killer key skills. - That's great advice. And I love especially how you imagine the corporate environment 'cause this is linked in a lot of our audience are people who work in a corporate environment. They may be in HR. They may be in various departments that, this is they're not data analyst they're not data scientist, they're not designers necessarily some of them are. And so that idea of being exposed to things outside of their own domain. when I do my workshops I'm showing things from The New York Times and some of the best work in the world and I think it is very helpful them for them to see that level of design even outside of what they may normally see and certainly do. - [Andy] That's right. I think you have to keep putting that Bill which you want to close on which is how we designers? I don't know if I consider myself a designer. I'm not a trained designer, but I'm making things. And I suppose therefore I am following a design process. So all those people out there who are working in corporate environments who think I'm not designer. You are 'cause you are making choices about how to lay things out. Colors, all these things we talked about. So that recognition can sometimes just open your eyes to new concepts and new ideas and new approaches. You don't have to use them all but still be inspired by what others were making out there. - Fantastic. I love it. That's perfect. I always say you're not only are you a designer 'cause you are designing but you are a communicator. In fact the most important thing that we all do is communications and sounds great. So listen, thank you very much for joining me Andy. I really appreciate our conversation and your insights, and I know that the second edition of your book just came out must be very exciting and everyone should check it out as well as checking out in his blog. He's a lot of great advice and that series of posts is really fantastic. So, thank you very much again. - Thank you Bill. My pleasure.