How To Learn Data Visualization (with Andy Kirk)

Yeah, let's kick it off. Enrico, you have to do it. You're the only one who can open this podcast. There's no one else. I'm sorry.

Hi, everyone. This is Enrico from data stories number five. And I'm together as usual with Moritz.

Hello.

And we also have again, Andy Kirk with us.

Hello. Good evening.

How are you?

I'm good. Just sat here looking at the rain outside, which is, you know, typical of England in April. So we're not as fortunate as some who can sit next to pools in Miami. But we're okay.

I've already trained from time to time there.

Exactly.

Yeah, it's much too hot.

Yeah. Moritz, what happened?

I'm good. I'm super tired. But who cares? We have a podcast to do.

The listeners don't care.

Yeah, they won't notice when I fall asleep. No, I'm doing good. Exciting times. Exciting times.

It's always exciting times for biz.

Yeah. But now it's especially exciting times.

Super exciting.

So today we are going to talk a hot topic, which is how to learn visualization. Lots of people are curious about that, and that's the main reason why we have Andy again with us, because he has some very nice training courses and he's a real expert on the topic. And, yeah, I think we wanted to start by saying that visualization is notoriously quite hard to learn because it's very multidisciplinary and lots of people really don't know how to approach it. Andy, what's your experience with that?

Yeah, I mean, what you say is correct about the multidisciplinary. I mean, the courses I do are introduction courses at the moment, and this is people typically who. I mean, this is the case for everybody these days. Everyone's touched in some way, shape or form by responsibility for data and presentation of data and communication of data. And the people coming to the courses represent the full spectrum of industries and organisation types, but also age groups and even skill sets. So you've got people there who perhaps are analysts already looking to top up their skills, or you might have academics or graphic designers looking to perhaps widen and broaden their portfolio skills. So everybody's kind of interested right now in the, I guess, the potential of visualisation in their own worlds. And it's just something that's really pleasing for me to see all these different types of people emerging on the scene and coming along to the course has got exactly what we were all expecting three, four years ago, when the subject really moved into the mainstream. I think the challenges we'll discuss in a second the challenge of picking up all these diverse skills. And this kind of convergence of all these traditionally diverse skills is something that's a challenge not just for people to pick up, but also when you're trying to conceive and structure and sequence a training course, especially in a single day. I mean, we're talking about maybe five, six contact hours in a day. There's no way a person can leave those 6 hours as an expert. And certainly my objective really is to give people a route into the subjects and give them the equipment through which they can discover the subject beyond the timeframe of a single day, because it's just impossible.

Yeah, I mean, what you say is correct about the multidisciplinary. I mean, the courses I do are introduction courses at the moment, and this is people typically who. I mean, this is the case for everybody these days. Everyone's touched in some way, shape or form by responsibility for data and presentation of data and communication of data. And the people coming to the courses represent the full spectrum of industries and organisation types, but also age groups and even skill sets. So you've got people there who perhaps are analysts already looking to top up their skills, or you might have academics or graphic designers looking to perhaps widen and broaden their portfolio skills. So everybody's kind of interested right now in the, I guess, the potential of visualisation in their own worlds. And it's just something that's really pleasing for me to see all these different types of people emerging on the scene and coming along to the course has got exactly what we were all expecting three, four years ago, when the subject really moved into the mainstream. I think the challenges we'll discuss in a second the challenge of picking up all these diverse skills. And this kind of convergence of all these traditionally diverse skills is something that's a challenge not just for people to pick up, but also when you're trying to conceive and structure and sequence a training course, especially in a single day. I mean, we're talking about maybe five, six contact hours in a day. There's no way a person can leave those 6 hours as an expert. And certainly my objective really is to give people a route into the subjects and give them the equipment through which they can discover the subject beyond the timeframe of a single day, because it's just impossible.

It's really interesting because I mean, I started my whole digital media thing like 1012 maybe years ago, and there multimedia was the big thing and everybody was learning sound editing, video, three de HTML programming director and everybody was like, oh, it's also new and everybody's just figuring it out and you have to know everything and learn everything at once. And now this whole scene has much more specialized. So yeah, everybody knows much better what type of multimedia is doing. But now the same thing starts in infovis. I have the feeling that everybody's learning everything and combining everything. I'm looking forward to see in ten years how the specializations.

That's right. And I think an interesting thing for me is people tend not to just purely be coming to the courses in the capacity of having a design responsibility. A lot of the people have got a role to commission work or to evaluate others work, and what they're trying to do is inform themselves better about how to evaluate and understand how to judge somebody else's work. And also you've got people who exist in teams, so you might have a single lead designer, but then they've got colleagues who are computer scientists or technologists or video editors. People are surrounded by teams of specialists and it's about how they harness the talent into a single cohesive project. It's not always people who are trying to pick up everything for themselves, but certainly have an awareness of what is involved and what is required to do this subject well. So it's such a fascinating field. I mean, you know, as things stand there, I guess there are very few of us doing the training offering at the moment. And that in itself sort of reflects that. This is still an emerging, evolving field, difficult to nail down and it doesn't sit still. So maybe in three to five years there'll be thousands of trainers out there and I'll be looking for a new job but we'll see.

I don't think so. There will always be something in this field to train.

What I think is really interesting is that there are so many areas that a person should be able to cover. You have to know about computers, humans, data, the design process itself is really, really broad. I'm wondering, Andy, from your experience, the people who attend your courses, what is the area that they have most problems with that they have a hard time to cover?

Yeah.

What's the biggest gap?

Well, I think the first thing to say is that I don't attempt to challenge to tackle a lot of the technology in the tool side, because I do explain to people that there's such an inconsistency of people's capabilities as they come to the sessions. There's no point me trying to find a session that talks about tools or programming in such depth that we'll be able to bring everybody along at the same time. So I kind of make that clear in the outset. But I think the difficulty people have, which is why I've responded with the structure of training that I have done, is the tactics is really, you know, when they are commissioned or they're briefed or they're asked by a boss or someone to do a visualization piece, it's typically surrounded by limited time scales or limited resources. And so they've got to find an efficient and effective way through this, you know, essentially quite a difficult creative and scientific task. And so what I try and do is give, give people this route through the design process. And obviously, it's very iterative and they have to revisit stages. But that, to me, is what people are seeking above all else. A set of tactics or a rough process or methodology that can be portable to any problem context, whether it's a small infographic or a large video installation, hopefully that will then give them the more efficient and effective way to tackle the projects. And obviously, everyone's looking for the silver bullet when it comes to what is the best method for this given scenario. But once again, there's so many variables there, it's difficult to say, oh, you always use this. And I think it's important also not to focus on hard rules or fixed principles as such because of that very reason. There are so many variables and different scenarios that are involved. And I recall, I think it's sort of a couple of months ago now, but there was a really interesting Twitter discussion between, I think it was Wes Grubbs and Hugh Moritz, and I think Jeff Clark was involved and Nathan, and it was about this balancing between. Do you give people principles and a kind of set of guidelines that they should follow, or should people just play and discover things? They're much more kind of serendipitous process. And that was a really interesting discussion, I thought.

Yeah, I personally think this is a huge theme. I mean, how much you want to balance the theoretical part with practice. And I cannot really see any of the two separate. I cannot see these two things separated. They really, in a way, they help each other. You need, of course, some basic knowledge and theory, because otherwise, basically, what I like to say is that if you don't have any knowledge to guide your design, you are basically playing darts. And that's really bad. And I see it all the times.

You can hurt yourself. Right. No, but I totally agree. I'm also. But I mean, that's also philosophical point, really. But I also think, like, you should first learn the rules and understand the conventions, you know, and understand how it works, basically. And then you can start to play with that and become creative and also play with these conventions, you know, and use them in interesting ways. But you have to be aware of all this body of work that's, you know, that's there already. And other people say, like, that's just, you know, that spoils you, and you should be as, let's say, as your mind should be as untouched as possible, then it's at its best creative point. I mean, I can see that in my kids, of course. I mean, my son is painting like crazy. Yeah. Better than I ever could, basically. So there's something to that, too.

Sorry, I was just saying it's the same in any creative discipline. I mean, you mentioned painting that when I did art at school, it was very much about painting between the lines and not going outside the lines. And then as you develop more confidence, you then start to become more expressive. Even things like football, you know, you can be so overly coached as a young kid that you actually lose that sense of.

You can overthink freedom.

Exactly. Yeah. So it's a real balance in that particular.

But the thing is, there is a strong functional component to data visualization. So, you know, it's not art, it's design. And we're building something that should work in some sense. And there you can get lucky by just, you know, fooling around. But usually it works much better if you know what you're doing.

Yeah.

Just saying.

Yes. Hashtag. Just saying. Hashtag.

Moritz. Are you aware that this is recorded? Yeah, it's carved in stone now.

Absolutely. I totally stand to that. No, I also think there is some, you know, some visualizations are objectively better than others. You know, I'm totally convinced of that. And some, I also think design is an objective activity where you can have objective measures. You know, it's not a matter of taste.

No.

What you do. Not really.

And I think serious, I mean, you.

Don't need to convince me on that.

Yeah, but some people, I mean, just saying, hashtag, hashtag, it's trending.

Then. Another important thing is we had a question from EJ Fox, who works at visually.

Then. Another important thing is we had a question from EJ Fox, who works at visually.

Oh, yes.

And the question he, or the point he was asking for a bit of a discussion about was in discussion with colleagues, it seems. Biggest gap for Dataviz newbies is needing a question before going further. And he asks, can you explain more about the importance of starting with a question or how to convey that process to newbs, newbies, new people? So I think that was followed up by a reference to a Ben Shneiderman mantra. And Moritz, you controversially retweeted and replied, did you say you want to discuss this because you don't agree?

Yeah, I mean, there's something to it. I mean, of course, you know, all these rules of thumb, of course they don't come from nowhere. So it's good if you have a question in mind. But I think it's such a cliche to say every infographic, every visualization needs a central question. I don't think it's true. And I think many great works just were produced from a fascination of a certain topic or a certain data set or a certain formal technique, you know, like.

But do you think so?

I'm fairly sure the great, great wind visualization Wattenberger and Viegas did, I'm fairly sure it just came from this fascination for the nice particle streams. And then they thought, ah, listen, you know what? We could use that for wind. Let's see if we can find a wind database, you know? And so I think that's totally valid. And there is no question behind this wind data visualization except the very obvious question, where does the wind blow? Yeah, but I mean, that's a topic for me, not a question. Right. That's true.

Also on another level, I mean, even from a more academic point of view, traditionally visualization has been used for exploratory data analysis.

Yeah.

And exploratory data analysis is, is basically looking at the data and seeing what can I see here?

And helps you find questions, not answers.

Right?

Yes.

And even searching for an hypothesis. So I think that's, that's true at many levels.

Yeah. And am I right, Enrico? And saying that's the distinction between, is it inductive and deductive research methods? Yeah. So one's boring.

I'm getting boring here.

We're getting too academic. Let's bring it back. So I think the response to EJ is that, of course many things will start with a question and start with a curiosity about something, but it's such an iterative process. You'll maybe start with curiosity, you'll embark on some exploitary analysis and then new questions, new discoveries will emerge and you kind of go on a cycle from there. So I don't think any lack of questions should prevent people from embarking on a database process. Just dive in there and see what patterns and relationships and interesting things emerge. But I guess, as you said, Moritz, at some point, maybe a topic will emerge, perhaps rather than a question.

Yeah, I mean, at some point it's.

Quite necessary forever and ever, but yeah, I think don't get too hot by needing a very specific, defined question. If it's there, great. But otherwise, no.

So I wanted to move on. I would like to ask you guys, if you would have to suggest to a novice who wants to learn how to do proper visualization, what is the basic skillset? What do you think are the most important things to start with in terms of knowledge? And then of course, we can discuss practice. Starting from studying visualization, what do you think is the most important body of knowledge that you need to acquire?

Well, do you want to tell us more?

Yeah, I mean, the basic thing you should know, which is often not reflected that much, is how visual variables work. So which visual properties of a mark can we vary to express something? And what different qualities do they have? And it takes quite a while to figure that out in detail. So because you can work with sizes, you can work with positions, with textures, with colors, with shapes, with rotation, you can combine all of that into glyphs. So there's lots of things to learn in this area. And basically, also, if you look at combinations, the learning never stops. And much is unknown there too. I think everybody who is in the field should know about the basics in this area. Also, basic perception rules.

The Cleveland's Miguel stuff.

Yeah, yeah, yeah. And because this is the thing we're working with, you know, it's like if you want to make music, you have to know about the tones, right?

Yeah.

So this is what we combine to bigger structures and to more complex things. And you, I think that's one thing I try to teach. When I teach, I'm not great at teaching. I should say that first. But when I teach, I try to.

And we have to lose Morich there.

Yeah, I try to make clear that this is, you should be really clear in how you use these visual variables. You should, you know, if somebody asks you, you should like, without hesitation, be able to explain how you did that in a certain way.

Exactly.

Because everything, there was a conscious decision. Nothing, absolutely nothing arbitrary about these choices.

That's an absolutely central tenet to my sessions, that everything has to be justified, reasoned, deliberate and explainable. And obviously, if you've got that knowledge of the visuals, great. But as you said, there are so many available combinations of things that, in essence, for new people, for new business. EJ Column it's almost a need for almost like a menu. This is the awareness of all these different fables that exist. And as you sort of move beyond the comfort zone of things like bars and lines into things like glyphs and motion and luminance and blur, it's very much really visual variables you're mentioning there.

Yummy. Missing texture and hatching, of course.

Yeah. But to offset the visual but variable stuff, obviously things around statistics and analytics and that kind of angle is obviously a foundation of the subject field. I mean, how many times do we see poorly applied statistics completely undermine a visualization piece? So I guess that's the, that's the stern offset to the design variables that exist. And then there's, I mean, there's things such as, I guess, taking the sort of visual variable angle a bit further. The Gestalt laws are obviously very important. And for those who don't know about Gestalt laws, these are a set of, I guess, laws that explain how the eye and the brain perceive shapes and patterns and relationships between shapes and patterns. That once again are the kind of foundation of how visualization succeeds, how we understand whether the continuity of a line is the same series or is a new series, and how you understand the different clusters and shapes that form from different position and relationships fail. Yeah.

I mean, visual variables, they describe quite well how a single mark works.

That's right.

You know, how large do I make that circle or where position it? And gestalt laws, they help you a lot in like talking about groups of markers and their interrelation, you know, and again, that's really, that's really important, basic knowledge there. Yeah, absolutely.

And I suppose picking up on the issue of questions and topics or curiosities, I think the best designers also have that sense, I guess, of a, almost a journalistic sense of what's an interesting angle. What's an interesting story to dig out? I mean, if we take someone like David McCandless, for example, I mean, David's core strength for me, he's got a really sound sense of what's an interesting subject, an interesting story to pull out from data. So I think that really adds, I guess, to the statistical side and going through the exploitative process, you could be on an infinite process of combinations and permutations of variables. But if you have got a sense of what you believe is an interesting angle, then that does help cut through a lot of the wasted effort. What would you say, Enrico, from your academic perspective, do you think is a real key ingredient?

I definitely agree that visual variables are the building blocks of visualization. And if you don't learn how to play with them first, the rest cannot really come normally. And there is one thing related to visual variables that I've experienced over and over, is, I don't know if you have the same opinion as me, is that particularly the way being trained yourself to use these visual variables in different settings, different problems, especially trying to. To have to come up with different designs on top of the same data, it's really, really useful. And among the visual variables, I think that position is clearly the most important one.

Oh, absolutely.

And I have seen people playing a lot with different colors, sizes and all the rest, but very few people realize that the most important thing is playing with layout. I think layout is really, really important. And the more you learn how to come up with different layouts, the more you have a whole lot of opportunities in front of you. Yeah, because this is what really changes the angle. You are talking about the way you use position to represent data. And then of course, on top of that, you can use all the other variables. But position is really, really important and I never see it stressed enough. I mean, I think it's really, really important.

Yeah. My 6th word advice is position is everything. Color is difficult in this area. And there's also, in design, there's usually a saying, get it right in black and white. So, you know, get it to work in. If it works on a napkin or like, you know, in black and white, you can be fairly sure that it will also work in a more refined format. But yeah, if you have the sort of the size relations and the. And the positional, you know, relations nailed, then you're almost there.

And that leads on to colour and the knowledge of color theory. And there was an interesting case this last week where I think it was probably myself, Moritz and Gregor were lazing with Simon Rogers about the guardian choropleth map and color scheme there. And, you know, there is, there are so many nuances to color that you can't just get away with throwing any color together. It needs to be so deliberate and so well informed about the different schemes, whether it's diverging, converging, whether it's sequential. I mean, you know, color spaces.

Yeah, I mean, there's.

It's color deficiencies, metaphors. Come on. I mean, for me, I think obviously everyone should know hopefully, about color. Brewer as colourbrewer two.org as a resource. But in terms of books, Colin Ware would be the go to guy for me in terms of color theory and understanding how the brain and eye works. But I think that's a key thing.

We will do a specific episode only on color. I can promise that.

Okay.

Because, I mean, you could do like ten different episodes on color and would still be frustrating.

Will this be like the Beatles albums? We'll do a black series and a blue series. And then obviously you need in terms of knowledge to clearly move on to things such as tools and software and programming to actually execute the work. And I mean, you know, I mean, this is probably once again another complete episode in its own right. But you've got to have an array of skills in, I guess, statistical packages and program environments and things like illustrator, Photoshop and publishing tools. So the great shame, I guess, of the field at the moment is just this fragmentation of all the tools that are out there, and there isn't one single tool that does everything. And I know that's a great frustration for people who come on my course is the sense that, oh God, I've got to learn six or seven different applications here just to complete a single piece. And, you know, many times, you know, it's quite overwhelming for many people, but ultimately that's the, unless you're doing something in pen and paper or crayons or tapestries or any kind of non computer related media. But it is ultimately what makes the think I'm alive.

Yeah, yeah. What I think is also important in terms of knowledge, which actually builds on top of that is having a clear understanding of what different kind of visualization techniques are available according to the data format you have in front of you. That's also really important.

Okay, so things like comparing time or those methods.

Yeah, if you have time, geo multidimensional graphs, hierarchies, having at least a toolbox of, I don't know, different standard approaches that you can try out right away as soon as you know what kind of data you have.

Absolutely, yeah.

And always a repertoire. So if you see a network data, you should not only think, I could make this network map, but you should also think matrix or radio layout. Right. So exactly. What I find really important is, apart from the data and the technic side, is this idea of to have a repertoire of metaphors as well. So there's a great book by Lakov from the eighties. It's called metaphors we live by. And he looks at language, I mean, it's mostly western thinking and English, but I mean, our whole language is so dominated by images, you know, that express something like, more is your language and less is down. You know, that's, in a way it's arbitrary. In another way it's clear, because, you know, it comes from the physical world also. Or that a stronger plant, you know, is bigger or like rises higher up. So that must be better. Some higher. And so all these directions, for instance, they have a deep meaning depending on which data you're using.

That's right.

Or in which domain you're operating. Right.

And there's a great book by Connie Malamud, the visual language for designers. I love it, which is great for that sort of knowledge as well.

Cool. Yeah. And I think it's something people often tend to forget who are very focused on the mathematical side of things, is that these things have an innate. They come with connotations.

Yeah.

Colors, directions, any type of graphic always comes with connotations and evokes certain metaphors or certain mental imagery.

Another. Oh, sorry, another small thing.

Your turn. Go on.

Okay, you both, it's my turn.

I guess we've eliminated each other.

Yeah, exactly.

Go on. Emerica.

Minus one times minus one is plus one.

So another thing I want to mention is that interaction plays a large role as well, or may play a large role. Yeah, I must say that interaction is a little bit of a Cinderella of visualization.

I thought it was color. What is color then? Damn it, I'm all confused now. Oh, God.

I mean, interaction is even more Cinderella than color.

Maybe it's the black art. Is it the ninja? Is it the black ninja? Maybe, yeah.

But I mean, what is really surprising from an academic point of view is that especially at the beginning of the early years of infovis, there was a lot of development in the area of very fancy interactive techniques to deal with visualization, and most of them have never been applied in industry or design or whatever. I mean, it's really strange. There is a whole set of interactive techniques that have been developed, developed during the nineties and they are basically dead. They are just in a few papers and nobody really cares about them. But I think it's an important aspect of visualization as well, especially when you want to do something that is a little bit more complex than only representing the data. As soon as you want to give the possibility to switch from one view to another or to, I don't know, get more information because you cannot really represent everything visually. I think interaction plays a very, very.

Large role and it has to be used really well. I mean, I had that with students often that they want to do either an animation or something interactive, you know, just because it sounds cooler or. Yeah, it could be more fun potentially. But if you do something, if you make something interactive that wasn't meant to be interactive, it's always like it's big fail. Yeah.

I mean, even.

Should we drop Shneiderman? Let's quickly schnup Shneiderman's mantra, so just go ahead. Yeah, so that's the sort of, he formulated this basic idea that any exploratory information visualization should first start with an overview. Then you can zoom and or filter and you provide details on demand. And if you think about it, I mean, 99% of all interactive info works like that, in fact. And so I think he pretty much nailed it. The question is how helpful it is to know that. But it's a good observation.

Absolutely. Yeah. And I think the, I guess the other thing that brings everything together is you've ultimately got to have a certain flair for design, an eye for design, an eye for layout and placement and arrangement, and just that's something about how you bring all this together into a single piece and that kind of design discipline as well. Knowing when to finish, knowing when have you got to the point of, you know, there's nothing more to do on this piece, but how do you learn?

How do you teach?

Well, that's it. I mean, that's something which I'd argue is perhaps more about practice than it's about teaching. Absolutely. But I think the thing to concentrate on the teaching is to make people aware that there is a need to consciously call something as finished and understand why that's the case.

That's the first. Yeah, that's true. That's the first step to, or to see maybe when something's overdone, under done and understand.

Exactly. Yeah. And I just think as well that the, I mean, design something that, which is very difficult to teach in its own sense, but all these other elements that we've discussed around visual variables about laws, man shows about color. That will give you the building blocks of a design capability, and then the rest is practice. I mean, that's the unfortunate thing. You can't just pick these things up individually and expect to run away with something successfully straightaway.

Yeah, but let's say somebody comes to you, he's in the beginning, or she of her twenties, and says, oh, I'm so keen on learning database. What should I do? So what's the first thing you should.

Well, you immediately go onto my website and book a course.

Book all of the courses there that was prepared.

But I mean, if we're talking about different channels through which you could learn.

Stupid questions.

Stupid questions. The wrong guy. Exactly. But I mean, I mean, certainly from my point of view, I was never, I was never taught these things. I, I was very much self taught and went through discovery kind of all my own. And it's, you immerse yourself in the books, you immerse yourself in the, the best websites and the practitioners and you keep on top of the.

Yeah, but would you say, go study, or would you say, go join a company? Or would you say go try it out yourself?

I think try out yourself in the first instance.

Yeah, yeah.

I think that personal void through this is probably more important. I mean, it depends on the company.

And get in touch on the web with like the usual suspects. I mean.

Yeah.

See who you can pick up a Twitter fight with and take it from there.

Yeah, exactly. But I mean, I think for many people, I mean, this is certainly true of myself. You know, coming from a background which was quite, perhaps hard statistics, information management, I was very much on the kind of corporate, shall we say, the Stephen Few style side of the world. And it does take quite a concerted effort to kind of release yourself from those shackles in many ways. And I would also equate it to art once again. You know, it's difficult to have that confidence to go outside the lines and to become quite expressive. But, and that's where I think the working on yourself by yourself is so important. It's much more liberating. There's no rules there, there's no company organization bosses going to say, well, you didn't do that. Right. This is going to do by next Monday. So I think, you know, that kind of personal practice. And as you said.

But will you learn alone, like how to all these design principles, I mean, so I would say take at least to an internship or maybe study design, because just to learn this process of sketching or understanding, like solving the wicked problems, knowing when to stop and these things.

Yeah, absolutely.

So. But maybe an internship is good, too.

So. But maybe an internship is good, too.

I also think. I mean, you've talked in previous episodes about contests, and, you know, if. If you're. If you've got the capacity and the time and the. The will to do it, there's absolutely no harm in taking those contests and the data sets that are on offer there and the briefs that come with them and throwing something together, submitting it, and then learning from what everybody else is doing. I mean, design is all about choices, and as you'll see on any visualization contest gallery, in so many different interpretations. And you can learn from those different interpretations, see what other people have discovered as a potential solution, and form your own judgments, whether that seems to work for you or not.

I mean, I would even say if you have time studying cognitive science hasn't hurt. For me, that was a good basis, too. I mean, it's a bit like a very long road, because first you study something where, I mean, technically, I don't use 80% of the things I've learned, you know, but it sort of provides that a lot of formal background in knowledge representation, computing, artificial intelligence, you know, and these things. Absolutely. Linguistics. So I think from a sort of coming from the second angle, not the directly applicable one, but the more the background, that's a great thing to do.

And everything that we say is a kind of. It's revealing that there is so little formal education on these things.

Yeah, there's none. I mean, there's nothing otherwise, just say, yeah, take that course and you're done.

Right, exactly. You know, or see math through to a certain level. There's nothing. I mean, and obviously, it's difficult. As we've mentioned, it's so interdisciplinary. Where would you pitch it? Where would you position this? And I've talked about this in previous blog posts on my site about where and how could you teach youngsters, you know, age ten, 1112, at the start of secondary school? How could you get the skills across at that early age? So it's difficult, and that's why people are now reacting afterwards, almost, to build up these skills off their own back in many ways.

Sure.

Yeah. I think what is really interesting is that if you like data visualization and you are in the position of starting a new bachelor or master, it's not easy to understand what kind of faculty you would have to go to. And there are so many options. So historically, you can do visualization in either computer science with a more. With a bend towards human computer interaction, or computer graphics. But you have lots of people coming from statistics, from design itself, from cognitive science, as you said, Moritz, like you. And there is a very long tradition in geography as well.

Absolutely.

Oh, yes. Yeah. Not to forget.

Not to forget. So basically, I mean, it's really hard to decide. I think, of course you have to follow your heart, but if your main goal is to learn visualization, and you already know that this is the thing you like, I really don't know what's best. I mean, probably I would say computer science because I'm a computer scientist, but any other faculty would work as well.

And I guess it also depends on where you've come from. So you might have a statistician who's looking to enhance their kind of design capabilities. So they'll already have that grounded. You might have other people coming from a graphic design, so they've got the kind of execution skills, they've got the design skills, maybe even the color theory, and they'll need to address all the gaps. So I guess it is the original perspective that you're coming from, the gap you're trying to bridge. But, yeah, I mean, all those, all those disciplines are ideal. But I mean, the interesting thing for me is if you're studying something like biological sciences or medicine or anything that's got a quantitative element, how do those people skill themselves in this field? I would love to see elective modules in universities that, yeah, everybody should have.

A course on that.

Everyone should have an alert, everyone can drop in.

I might be giving a course to like 1012 PhD students from all kinds of fields. So I'm just discussing that. And this will be fantastic because they all do visualization. They're just not reflecting it, you know, that this is something or how much you can do there. I think there's. And for me, it will be totally interesting to look at their research, too. So it should be fun.

Yeah.

I mean, in the end, visualization is a basic tool, and you can imagine.

In an ideal, it's like writing or, you know.

Yeah, exactly. This is what I was saying. In an ideal world, you would have to, you would love to teach visualization to in the high school as well. I mean, why not?

Yeah, absolutely. It didn't go any further, unfortunately, but I had a teacher from a local school approached me about an interest that he had in, I think it was his maths class, and he was looking to find out, unfortunately, didn't go any further for various reasons. But I mean, I guess let's make this a public appeal. I'd be very interested in getting together with any teachers out there are interested in. I don't know whether it's a maths class or any class, really. I'd love to find the technique to teach at that level now. You know, I'm so far away now from school, I can't remember really how or what is the most effective technique at that age to impart such things and how deep should you go? But I'd love to have a go and see what we can do. Because if you can get people equipped at that age, then wonderful for them.

So for self teaching, I mean, what would you recommend? Are there any good lectures, books? How would you go about.

So some time ago, I wrote a blog post on books and other resources. And I think, Andy, you had something similar in the past, right?

Yeah, I copied immediately after.

The usual thing.

Yeah.

Plagiarism, french thumbnail.

It's sickening. That's right.

I remember what I tried to do is that there are so many resources around that it's really, really hard to decide what to get right, and especially if you go on the web, but even if you just want to focus on books, there are so many right now that it's really hard to know where to start. So I have a few books that I think are really, really basic and they do not overlap too much. And for me, these books are, and I can tell you why I really like them. The first one is show me the numbers by Stephen Few. I learned so much from this book because it's basically only about very basic graphs, like bar charts, timelines, even tables and so on. But it helps you to understand how, what are the basic building blocks of visualization? If you don't master this first, it's really, really hard to do anything fancier that is also effective at the same time. So I think this is a really great introductory book from my point of view.

Yeah, that was my first book.

Yeah. Okay, great.

There you go.

I never read it.

Oh, yeah, get off the line.

Yeah.

The second book I read only books where I appear somewhere.

Yes, Cinderella, he doesn't even kill the book himself.

So the second one is more academic is readings in information visualization, using vision to think. And that's a key.

I mean, but it's a bit dated, isn't it funny to say?

It is, it is. It's a bit dated. But what I normally suggest is to start to read chapter one. And I think chapter one still is the best introduction to information visualization ever. Ever. You cannot go wrong with the first chapter there are so many basic principles summarized in a few pages that, I mean, it's really, it's the best you can get. If you want to start with something small, and it's a very solid foundation from the theoretical point of view, I think that's the best you can get.

Okay, could be, but maybe if I will have to see how. I think it was fairly expensive and.

You're not in it.

And just for that one chapter, and I mean, a big part of the book are classic papers, but I'm not sure how much they would help somebody who is now getting started.

You know, it's a collection of paper, the rest is a collection of papers. And most, it's mostly outdated. Yeah, but the first, if they could publish only the first chapter, I think it would be really, really valuable.

Yeah, could be.

And then, of course, you cannot not read at least one Tufte's book.

That's true.

And I think the visual display of quantitative information is still the best. I mean, you cannot go wrong with that one.

Yeah, that's true.

It's the most basic one. Basic principles, still not easy to read. I think every Tufte book is not an easy read at all.

And it's more, I guess, almost like a stream of conscience, isn't it? It's a set of ideas. I mean, you would, you wouldn't have that next to your computer and follow step by step what to do, but it's a reference for sticking a bookmark in different pages. This is a really great design concept that will transcend all tasks and just stick in your mind as a thing to remember and to recall as you go through the process. That's how I've used that book. It's a great book.

Then the last one I would mention is information visualization, perception for design from calling Ware, which you mentioned already before in the podcast. Yeah, and that's the, I mean, if you want to have, and you have to have a solid foundation from the perceptual point of view, I mean, that's the book. If you really want to choose one book, that's the book. If you want to get serious about visualization, you have to read this book. And I have a small criticism about this book, or a kind of warning. In many cases, it's very hard. At least I found it hard to translate the content of this book in practical rules.

Exactly. Yeah. It's great to learn about all this background info, you know, and at some point then you go, aha, you know, during workouts and think back to it, but it's not really a guide. You know how to. How to do it, right?

Yeah, yeah. It's not a guide, but it's important. And you get the feeling that once you know these things, everything is much more. Whatever you do in visualization is much more well reasoned. And you get. You acquired those skills that help you justifying the. Why you do things in a certain way.

Yeah, yeah. And I'm just looking through the books that I put up on the, on the site and it's such a, it's such a process of layering up your, your knowledge because I mean, you get the foundation stuff there. But I mean, you could also argue that there are elements of, as we've mentioned before, William Cleveland's book, visualising data, or even actually, I probably don't bother getting the book and just try and get. Track down the research paper that he did around 83, 84 with McGill.

Yeah.

And then, you know, there's different perspectives that you can drop into. I've always got a lot out of books like Donald Norman's book and Bill Buxton, Alan McReach. And, you know, all these different perspectives that just reinforce messages or just bring a different kind of angle to things. Merlin Wilkinson Grammar, graphics. I mean, there are so many books out there. It's a wonderful kind of subject to get into if you've got an appetite for reading and consuming books.

I must say that what is really interesting is that recently we also saw so new books came that are more about data visualization, that they have more practical angle.

Yeah.

And I think that's. That's great, too. I mean, the book from Nathan Yao.

Yeah, that's.

That's a great book. And also the other one, Noah, is it called please help me? Yeah.

Design data visualizations.

Yeah.

Yes.

These are two books that are much more practical, and I think they very nicely complement the more theoretical ones. And they are equally important.

Is it too shameless if I announce that I'm writing a book as well?

Wow, that's a scoop, Moritz. We have a scoop.

Yes. Yes. I'm working on a book. Excellent.

Wow.

In fact, I'm potentially working on three.

What will it be about? Soccer.

It's about British soccer.

Microbreweries.

It's. Yeah, it's color in between the lines, an introduction guide. But I guess it's going to be, in some respects similar to the sort of stuff that I do basically training around. So I'm going to leave it quite enigmatic and elusive and I'm just going to leave it there. But you know, I've plugged it and I'll keep people informed as things develop. But, yeah, it's an interesting challenge, actually translating knowledge into training and then training into a book format, because it's a completely different format.

And on training, you rely on the dialogue as well, interaction. And in the book, you're like talking to the ball.

That's it. And it's linear and it's so difficult. Yeah. Watch this space.

Anyway, but publisher or are you like.

Yeah, I've got a publisher.

Fantastic. Wow, wow, wow.

We have a scoop.

Gonna be the best book ever.

Oh, my word.

Does it have pictures?

It will have pictures.

That's fantastic.

Not couple pictures. It'll have pictures. But I think. Is it going. Go ahead. Sorry.

I hope it's not going to be another one of those table books that you never read.

A coffee table book? No.

Coffee table book.

Yeah, no, it's actually, I think, going to be published as an ebook to start with. Okay.

Okay, great. Does it have an animal on the.

COVID No, there's no animals or crustaceans or any version of it.

Do you want me and Moritz on the COVID On the back cover, maybe?

On the back cover, yeah. Peer reviewed people. Yeah.

Very small.

So. But, yeah, so, yeah, I mean, going back to the. This section, I mean, books is something that you can, you know, get from libraries and dive into and, you know, I think just not just myself and you, Enrico, I think George Georgie Camos did a post as well. So there are a lot of people out there sharing their experiences of what books have been important to them. So, yeah, it just goes back to the point about immersing yourself in the subject.

One thing that helped me a lot was I was at some point just googling for information visualization lecture PDF. And it's amazing, you know, because everybody puts their slides online. And so I went through Jeff Heer's slide, you know, and all some really good lecturers, this one.

Brilliant.

And then just followed up the references from there. So that was a great summary of what I should probably know. And, you know, and the other thing is on iTunes, you just Google for visualization. There's some complete visualization lectures online.

Yeah.

Some from Harvard, some from Stanford.

Yes, that's what it's saying.

I mean, it's a bit tedious to watch a university lecture that wasn't meant for video. But still, you can take a lot out of these and then you can.

Go online and get a full certificate and say you've passed a degree.

Yeah, they know they're now doing these math lectures for like hundreds of thousands of people. I mean, no info with one, but it might come.

It must come at some point.

Yeah, it must definitely come. So that's an amazing development, if you think about it.

Yeah.

Gotta love the Internet.

It's headstair.

I think it's gonna go.

It's got a future.

And I think Tamara Munzner is going to write, publish a book.

She's writing on a book. Yeah, I've seen her draft.

And Alberto Cairo's got a book coming out. It's already out. He's translated into English. And there's some other people, and I can't remember who they are. Now, I didn't write about it, but it's. Yeah, it's a rich period for good quality books.

So before we finish, do we want to mention a bit about how to learn the practice of visualization? I mean, I don't know. We have been discussing a little bit about how to acquire the basic knowledge, but let's assume that now you have this knowledge, how do you translate it into actually being able to design visualizations? Because it doesn't really come naturally from them. I mean, from my experience, I noticed that it's almost impossible to study visualization for hours and then directly translate what you learned into something that is really valuable. It's more of a craft.

So before we finish, do we want to mention a bit about how to learn the practice of visualization? I mean, I don't know. We have been discussing a little bit about how to acquire the basic knowledge, but let's assume that now you have this knowledge, how do you translate it into actually being able to design visualizations? Because it doesn't really come naturally from them. I mean, from my experience, I noticed that it's almost impossible to study visualization for hours and then directly translate what you learned into something that is really valuable. It's more of a craft.

Right.

You really need to do it over and over to learn how to do it.

Well, yeah, I think you need to learn some programming language or some. You need to have some environment that is fairly, that's more than template based, you know. So I think you cannot just use Excel and try and use Excel. There's creative uses of Excel. Anyhow, I think you need to pick, just be it HTML or processing or whatever environment.

Actually, that's a good point, because we got another question from Lane Harrison, and he's a PhD student at UNC Charlotte. And he asked the question, which was, if I can find it, so what the pros and cons of DIY tools such as D3 processing versus system tools such as Tableau. So that's question to you, Moritz. Let's put it to you first.

Yeah, I mean, in a way, it's quite obvious. I mean, of course the boundaries are blurring, so every tool brings its assumption with it. Even D3 is not totally unconstrained. Right. But it's a bit more open what you can do there than, for instance, in Tableau. And of course, Tableau is a bit like on rails, right? There are certain things you cannot even do with certain data types because they don't make any sense. Yeah, well, you can make totally nonsensical stuff in D3 if you like. So with great power comes great responsibility. That's quite a thread. I think in order to make a really good visualization, you have to have that freedom of make also nonsensical stuff. And then, and I think it comes back. Use it power wisely.

Yeah. And it comes back to capability in time.

Yeah. You know, and thinking about encoding again, I think the encoding you choose, you know, that makes the visualization in the end, and you have to explicitly formulate that. So sometimes in Tableau, I'm not really sure, like, is he scaling the circles, you know, with the square or is it linear? Hmm.

I don't know.

It's like, let's measure it. But if I program it, I know, like what's the process from data to visual? And I think you should know if.

You, it's the control.

It's like, you know, if you're a chef and you're cooking, you don't use like premade food, you know, it's like, hello, please. You chop your carrots and you peel your.

Can we suggest? So if a person wants to learn infobase, do you think there's no way to do it without knowing a little bit of programming?

Yeah, yeah, yeah, yeah, me too.

That's a recurring question and I agree. I mean, but on the other hand.

Actually, can I take my answer back? I want to take my answer back because, well, I guess this is visualization. Yeah, I suppose if we expanded things to include things like infographics, but obviously there's a great deal of infographics that are done through illustration and, you know, obviously built on some sort of data or information. But yeah, I guess visualization, I mean, to fully unleash the power of a visualization, then programming languages is unfortunately a requirement.

Yeah, I mean, information graphics has been around before computers, so I mean, you know.

Yeah, so that sort of makes sense. William Playfair managed all right.

But I think if you talk about data visualization, it's like, it's quite clear.

Yeah. But what I want to say at the same time is that you don't need to become a software engineer to do this.

No, you can totally like wing it.

I mean, you can start with processing. I think it's really nice because it's not a too steep learning curve there. And not necessarily you have to learn every detail about Java. You can start with processing and see what happens. Right.

Well, that's interesting because I need to learn, and I've kind of lined up this summer to focus on one particular tool. So you both would say processing has got the more accessible routing versus d.

Three because of the lack of UI toolkits.

Okay.

Yeah. And so I think if you really want to do interactive stuff, processing is great if you have just one view in your application, but if you have like three different tabs.

Yeah.

And different controls and zooms and pans and whatever. It's a lot of work actually. So I think flash is the perfect environment. I think it's a bit tragic that it's going away because it was really made for visualization. So I'm really bummed here, but I think for fresh people, I would say build it inside the browser you have. HTML is fantastic and CSS and JavaScript is okay, but there's lots of libraries and the community counts. But I mean, processing would be my second choice, especially if you go more high definition and installations and cinematic effects and so on.

Yeah.

And I think it's also a matter of the nice environment it provides.

I mean, there's a very active community, lots of source code around, lots of.

Source code, lots of documentation, and then you basically, you install it, you run it and you can start writing code right away.

Yeah, yeah, yeah. And it's nice people around it. So, yeah, everybody active in the processing scene is really cool. And that, you're right, that makes a big difference. Yeah. Good, okay, let's do that.

Let's do that. One last thing I wanted to mention about practice is another question that I always get is how do you find data and how do you.

Let's do that. One last thing I wanted to mention about practice is another question that I always get is how do you find data and how do you.

Yeah.

Find interesting problems. I think that's another big issue, especially for people who are just starting.

Big question.

Yeah. What do you think?

There's plenty of problems going around, but I mean there are so many and growing in number sources of data, data market info, chimps, guardian data blog, you know, I guess it's whether or not these are the datasets that you particularly want to have a go with. But if you are just looking for something to practice on, as we said before, the value of the contests, there's usually perfect data sets prepared for you there and also a problem or a brief that comes with it. So they're great test environments to practice. But I mean, the amount of data on the web these days is just infinite, so you shouldn't struggle to find World bank data.

Two things I find always interesting, or where people don't think about so much, but there's a lot of stuff on your hard drive. And also, let's say, skype, your chat history is on your hard drive, so you could totally visualize that APIs. The other thing is you can also generate your own data sets. And I like that for teaching to say we started zero, first we generate a dataset because then you have to think about, okay, what do I actually measure? How do I measure, you know, what's the granularity? Isn't it redundant if I measure this, but that as well? I could just write down this one value because the other one depends on it and stuff like that. So I think that's a good exercise too, to think about how could I generate my own data?

Yeah, I think that's a great advice, because if you have data that you care for first and you can really understand, then you have a huge advantage. It's a lot easier to create something that is meaningful and interesting and not, I don't know, obvious, right?

Yeah, exactly.

And I think just kind of closing the benefit of the practice angle down. You need to put your work out there. You need to let people evaluate it and critique it and feedback, because otherwise you kind of exist in that bubble of just you and the design and your own taste. And that's kind of good for only so far. So, you know, get a Tumblr blog, get a blog, get something going where you can publish the work and seek feedback from people in the field to kind of refine the skill. And that'll be where the practice actually comes alive. And that's where the experience will develop.

Yep, absolutely.

I think that criticism has a big role in learning the how to do good visualization, because by seeing what other people do and having a critical eye and trying to think about what works and what doesn't work, it's really, really useful, especially at the beginning. There is the blog, the y axis, which I really, really like because it's totally devoted to very detailed criticism and constructive criticism, by the way, of anything.

And also very balanced, right?

Yeah, very balanced. Very balanced. It's always mentioning good and bad things. And I think this is a perfect example of how you can acquire better skills in producing good visualizations, because first you have to understand what works and what doesn't work. And the best way is to look at what others do and also actively think about how would I redesign it in order to overcome these problems. And normally what you find is that as soon as you try to fix a problem, then you have another problem somewhere else. Right?

Then you have to.

Yeah, I mean, you fix this problem and you and you create another problem. So I think it's important to go through this process.

Yeah. And also, everything is a trade off. So in the beginning, you often think like there's the perfect solution, but usually there isn't. So to live with these trade offs and basically it's all about picking the right trade off, you know?

That's right.

The least worst solution.

And I was. I had an email conversation with Jay Thorpe about the process that he went through for recent work for, I can't remember the magazine. Escape my mind. Popular science.

Popular science.

138 years of popular science. And what was interesting about what he revealed there was he conceded that his process is riddled with errors and blind alleys and things that don't go away. So, you know, even, even the best.

It's a sad profession.

Yes. It's pathetic. It's built on failure. Don't expect to get things right the first time because it's an iterative process.

But I'm constantly cursing during work. I mean, during the day. Yeah. Usually it's not fun. I mean, for others it's fun. Exactly. Maybe I should do that, like the live stream.

Wordly.

Maybe, maybe not. Yeah. Damn it. Stupid thing. Yeah. Anyhow, I think we're good. I mean, if now people don't know how to learn it, what else can.

What hope is there?

Yeah. Not that hard. I mean, come on.

No, I think it's good to.

I mean, we just barely scratched the surface, but.

Yeah, yeah, but we have some blog posts around and we will keep talking about these things in the future.

I believe so, too.

Andy's book.

I'm so looking forward. I'm really curious now. Damn it. Wow.

I mean, it's.

Maybe. Maybe the listeners won't believe it, but we didn't really know it.

That was a genuine scope.

I didn't even know. You should write, basically.

I can write. Yeah.

Was it a present for us, Andy?

It was. It was the gift of being invited back to episode five.

Probably he just made it up and now he's, like, checking out publisher websites.

I'm just quickly writing about Google for.

Free online books publishing. Ghostwriter.

Quick insert author name. Yes, exactly.

And now pick a cover for your so called book.

It will come, it will arrive.

Don't worry. I'm totally looking forward to that. Good. Yeah.

More, Moritz?

Yes, I'm looking forward to c conference, by the way, which is in nine days or something. So if you're in Germany, pick a ticket. I think there are still tickets available. There will be Emmanuel Lima, Stephanie Posavec and on Sunday, we'll do a little workshop c with some of the speakers.

That's great.

Yeah.

One of my students is coming.

Oh, that's nice.

Yeah, that's nice.

There's some big names there. I understand. There's young Willem. There's Gregor on the workshop.

Yeah, everybody.

The who's who.

Benny, moon widow care. Yeah, yeah, yeah. It's gonna be great.

Paulo going?

No, I think he has some other obligations.

Okay. So. But it's a good. It's a good lineup. So if you're in the neighborhood, drop in.

Yeah. In the neighborhood means, like, around thousand miles in the flight there without changing flow.

That's the way I view Europe. It's just one big neighborhood. So that's training, is it not? That's the end of the training.

That's the way I view Europe. It's just one big neighborhood. So that's training, is it not? That's the end of the training.

It is. It is, most definitely.

So it was really interesting.

And, I mean, we could do this, I think, every week, basically, but, yeah. So much to learn. Yeah.

Yeah.

Okay, guys, so I think. I think we have to cut it here. We have been talking for three days.

It's always with the Andy. Always with the talking. Yeah, yeah, yeah.

It's a bonus. Bonus material. We'll publish the outtakes as well, shall we?

Bonus for one. Okay, guys. Yeah. So it was great to have you.

Thank you very much, Andy. Thanks a lot. I mean, it's great having you here.

We'll have you again when you have your so called book out, okay?

Yeah, of course.

Absolutely.

Okie doke. It was great having you.

Thank you. Plans for Moritz?

Thank you. Plans for Moritz?

Do we have a plan?

I think we have a plan for having the next episode on color.

The color show. Yeah. I hope we can do the next one. We want to have a very special guest on colors. But if we manage to do that, then the next one should be on color, I think. Yeah, because color is a good topic. Otherwise, we just make something up and go with it.

Eddie. Probably from Mister Tuft.

Eddie. Yeah. Why not? But he's not great in color. He's more like a black and white guy.

I'll tell you what you need for colour. You need Nigel blue.

Oh, good point. Yeah. At least, like, make one statement in between.

It might be a limited discussion, but.

Yeah, I mean, yeah, he can only talk about color, not colors, but color theory. Yeah, color theory. That might work with him. No, it might be an option, too. It's a good point. Yeah, we can see.

Okay.

Okay. Okay, guys, have a great time.

Enjoy your week. See you later. Bye.