Welcome back to the PolicyViz Podcast! I hope you had a great summer and got some time to rest and relax. I’m really excited for this season of the show–I’ve got some great guests lined up to talk about their work and how to do a better job communicating data.
I’m kicking off this season with my good friend Alberto Cairo. Alberto is a journalist and visualization designer, and the Knight Chair at the School of Communication of the University of Miami (UM). He is the author of several books, such as The Truthful Art and The Functional Art. He is also the director of the visualization program at UM’s Center for Computational Science. He has been head of information graphics at media publications in Spain and Brazil, and he currently consults with companies and institutions such as Google.
Alberto’s new book, How Charts Lie, will hit bookshelves any day and is sure to be a go-to resource about how to avoid creating and reading misleading charts and graphs. In this week’s episode, we talk about the book, why he wrote it, and how he hopes it will help people be better data and data visualization consumers.
Support the Show
This show is completely listener-supported. There are no ads on the show notes page or in the audio. If you would like to financially support the show, please check out my Patreon page, where just for a few bucks a month, you can get a sneak peek at guests, grab stickers, or even a podcast mug. Your support helps me cover audio editing services, transcription services, and more. You can also support the show by sharing it with others and reviewing it on iTunes or your favorite podcast provider.
Jon Schwabish: Hi everyone. Welcome back to the PolicyViz Podcast. I’m your host, Jon Schwabish. I hope you had a great summer, and I hope you had a chance to rest, to relax and spend some time with friends and family, and I hope you’re ready to listen to some more great interviews here on the PolicyViz Podcast. I’ve lined up, I think what is going to be a really terrific line-up of guests, uh, folks who are doing great work in the fields of data visualization and data science and presentation skills, uh, even some folks who are working with sound in data visualization. So I’m really excited about that, uh, to bring all of the guests your way. Um, before we get into this week’s episode, just a quick note that this show is, uh, supported by you, uh, the listener. And if you would like to help support the show, uh, either financially or just by spreading the word, that would be great. So if you’re interested in financially supporting the show, you can go over to my Patrion page and sign up and you know, send a few bucks my way so that I can help pay for the sound editing that occurs, the transcription services, all the stuff that’s needed to make this show, uh, what it is. If you’d like to support the show in other ways, I’d really appreciate it. Share it with your friends, put it on social media, uh, let other people know about it so that they can hear, uh, all the lessons and skills and tools that they need to do a good job and a better job with their data and with their analysis. So I’m really excited to launch this season with my good friend Alberto Cairo. Um, you probably know Alberto’s name if you’re in the data visualization field. Uh, Alberto is a journalist and visualization designer. Um, he is the night chair at the School of Communication at the University of Miami. Um, and he’s also the author of a new book, How Charts Lie that is coming out, uh, in just a couple of weeks. Um, Alberto spends a lot of time thinking about how people read visualizations and they perceive visualizations and how visualizations can mislead them. And that’s what this book, how charts lie is all about. So in this week’s episode, uh, Alberto and I talk about the book, why he wrote it, um, how some of the charts he had to describe and critique, uh, maybe made him a little angry, a little disappointed in the way people present their data. Um, and we talk about, uh, charts that are misleading intentionally versus those that just are using bad data visualization techniques. I think there’s a lot to learn both from the interview and from Alberto’s, uh, forthcoming book. So I hope you’ll enjoy this episode. I hope you’ll stick with me for the next few weeks to say, I’ll bring you some great guests doing, uh, work in the fields. And so here we go. Season number six of the PolicyViz Podcast is going to kick off with Alberto Cairo.
Hey Alberto. How is it going?
Alberto Cairo: Doing good, how are you?
JS: I’m good. I’m good. Getting towards the end of summer, kids getting ready for school, Middle School for my oldest. So there’s a lot of nervous going on over here.
JS: What about you?
AC: A lot of work. So I’m in the same situation, my kids are about to be in school, so we are getting ready for that. So it’s a busy time.
JS: Yeah. Yeah. And you’ve got to get ready for teaching in the fall, right?
AC: Exactly. And in the exact same week. So it’s like…
JS: Oh wow. Oh, you start early.
AC: And everybody goes back to school in the same week.
JS: Right, right. And you’ve got this new book, and you’ve been promoting it, and you’re going to be getting out there and doing the rounds. So, um, uh, I want to talk about the book and some other things that we can, we can talk about, but um, maybe you can just start by giving us the description of the book, why you wrote it, you know, give us, give us the pitch as it were.
AC: Sure. So the, uh, the title of the new book is How Charts Lie, getting smarter about visual information and it comes out in, in October the 15th, um, this year and I’ve been working on it for the last couple of years or so. And um, the reason why I wrote this book is that, um, before and after the 2016 election, I started getting interested in the many ways that people misinterpret, misuse, misread different kinds of data visualization and then how they employ them to push their personal agendas, they, they to push ideas to promote ideologies, etc. So I started doing some reading and you know, some, some writing and some thinking about what could be done to avoid that. How I could perhaps help, you know, non specialists, normal people, people in the general public become better readers of charts. So what things we should pay attention to whenever we see a data map or a data graph or a diagram, etc. What are the features that we need to read, things that we need to pay attention to, the most common ways in which a chat can mislead. And above all, how we can avoid lying to ourselves with the charts that we see by, by projecting what we already want to believe onto a chart that tells a completely different story, right. So I started getting some ideas about that, putting them on paper and that’s where the book comes from.
JS: Yeah, no, that last point is interesting, how we put our own biases on top of what we’re seeing.
AC: Yeah. Yeah. That’s a, that’s a, actually, then the, the most, the most common problem whenever we see, we see charts and one of the main reasons charts mislead us so often. So I think that is the first thing that we need to address if we want to become better readers of charts.
JS: Right. Let me ask, so I’ve seen a lot of these graphs that you’ve, you’ve critiqued and you’ve been doing a lot of, lot more blog writing, I’ve noticed to, to highlight some of these. Was collecting these and writing about them, was it like infinitely frustrating and infuriating?
AC: Um, not really. I mean there are, there are some examples in the book that are, that are very infuriating and, and worry and even sad there is there. For example, I talk about, um, a Dylann Roof, the guy who, who enter a Church in South Carolina and he shot… a racist who enter a church and shot several people. And by reading that, that very sad and very, you know, terrible story, I discovered that one of the reasons why his racism increase is that he read several reports by several racist publications that were misusing data and charts to sort of prove quotation marks in there that African Americans target whites more often than whites target African Americans when it comes to committing a crime or when white, when black criminals target white people, that happens more often than white criminals targeting black people. And I’m not going to get into the details of why, you know, these, all these graphics are, are basically um, crap. Uh, I describe that in a lot of detail in the book. Um, but it’s like it’s a very sad story and actually demonstrates that bad charts can sometimes have terrible consequences. So this guy Roof will be a racist regardless of the existence of these graphics I believe because he was a racist since he was a, since he was a child. But I think that the charts contributed to basically ground his beliefs and strengthen his beliefs even more. So cases like these are certainly infuriating and there are a few other charts in the book that I believe that were designed intentionally to mislead people. And I called them out obviously, but most of the examples in the book are examples of charts that are otherwise perfectly designed, but they are often misread or misinterpreted. And this is not infuriating. It’s just a fact of life. I mean, we are taught, you know, a or we are, we are told that, you know, we should be able to intuitively understand visualization that visualization is easy to read, that a picture is worth a thousand words and things like that. And in the book I tried to demonstrate that all these myths are actually myths and that we need to abandon them that visualization is sort of like a like written language. You need to pay attention to it. You need to read it carefully, uh, in order to interpret it, uh, in order to interpret.
JS: Yeah. Yeah. I’m always surprised when I show people, you know, a different type of graph. You know, like, you know what we, you know, like a slope chart or a dot plot that we have of, you know, that we in the field know now instinctively. And people say I can’t show this to my boss or my manager or whatever because they’ll never understand it. And I find that interesting because it’s not like we know instinctively in our DNA how to read a bar chart. We have to learn how to read a bar chart.
AC: Yeah. I, I perfectly remember when I learned to use a, to, to read a scatterplot for example, and it was not intuitive. I needed to pay attention to the chart. Take a look at the axis, um, read a little description, read a little caption that the chart had in order to interpret it correctly. So in the book I talk about visual literacy obviously. There is a term to refer to visualization literacy. The term is graphicacy. And I explained where, where that term comes from and other authors that have used it in the past. And I said that the problem is that we like graphicacy and but we can increase graphicacy. The, the problem with, you know, people who react negatively to novel graphic forms is that they say, well, my reader is not going to understand this chart. So they refrain from using that chart. But that’s the wrong response. The wrong response is to say, well, if this is the best chart to represent your data or to tell or to convey the message that you want to convey, do use it. Go ahead and use it, but also explain how to read it right to help guide your readers by the hand in order for them to understand what’s going on in that chart. And then the next time that they see that same type of chart, they would be able to read it on their own. You have, you would have increased their graphicacy.
JS: Yeah. Yeah, absolutely. So, so you talked about a couple of these examples where the graph was intentionally misleading or misrepresenting the data. What’s the balance between those type of charts that are intentionally misleading versus one that, that use bad data visualization techniques?
AC: Well, in the book itself, in How Charts Lie, I would say that 20, 25% of the examples are charts that I guess, it’s just a guess that I guess that are intentionally misleading. Um, and the other 75% are charts that are either well-designed, but misinterpreted anyway, or charts that are designed with good intentions in mind, but that they employ, you know, visualization techniques that are not appropriate for a particular audience. And as a consequence of that, they end up being misleading any way. The result is [indiscernible 00:10:46]. The audience, audience that is reading their graphic is misled. So I would say that that’s sort of the sort of the balance because again, I’m much more interested not in the intentions of the designers who create those charts. I’m much more interested in the consequences of those charts of how the public can use charts to basically, you know, have better lives to be more informed, to be better informed.
JS: But what, when you say that a graph is being misinterpreted, do you put the onus or the responsibility on the graph creator or the reader of the graph?
AC: Both. Actually this is a point, this is a point that I make in the, that I make in the conclusion of the book in which I say the first responsibility is on the designer. So the designer needs to make an effort to, you know, try to understand who the audience is going to be, try to guide the audience, blah, blah, blah, use, you know, appropriate visualization techniques. If the designer uses a novel graphic form, explain it so people will understand it. Right? So there’s obviously a responsibility on the part of the designer, but there is also a responsibility on the part of the reader. And this is, this is connected to what I said before about the myths that’s around data visualization. We have talked, we have been told so many times that our visualization needs to be intuitive and must be easy to read and simple, et cetera and we have been told that a picture is should be worth a thousand words and blah blah blah, that we have internalized that we can understand a visualization just by looking at it and rather than reading it. And we do need to read it. We need to pay attention. You cannot assume that you will understand that chart if you don’t read it carefully. You do need to read it carefully and if you don’t read it carefully and you misinterpret the chart, if the chart is well designed, then the responsibility is not the designer’s responsibility. It is your responsibility as a reader.
JS: Yeah. Where do you place the responsibility with using the data in this, I guess, process of extracting and analyzing the data, making the visualization and then publishing it? Where, you know, how do you separate the data part from the visualization part?
AC: Well, in most cases you cannot really separate the data from the visualization just because the visualization is sort of the data mirror made visible or the data made physical so people can see patterns and transcended data. So the data and the visualization are intrinsically connected unless, unless you’re doing this as just an aside, unless that you’re doing that data art project, something that is a little bit more expressive. In that case, the goal of the visualization is not to illuminate anything about the data. It’s more to create sort of an aesthetic experience based on the data. In that case, the data’s a little bit secondary in comparison to the visual experience, but in most cases when you do a visualization, the point or the goal of the visualization is to be able to see something from the data. Whose responsibility is it? Well, it’s the designer’s responsibility, obviously, to try to get the data right, try to talk to experts who know much more than you do about the data, etc., etc., to verify what it is that you are presenting to the extent of your knowledge. At the same time again, in connecting to what I said before, there is a responsibility on the part of the, on the part of the reader to try to, you know, read the graphic carefully and not make assumptions about the graphic. And this is a specific example that I have in the book that explains this idea well. I will show a chart and let me say beforehand that this is a, this is a mistake that I have made myself repeatedly. So there is a chart that shows, it’s a scatterplot that shows a positive association between cigarette consumption per capita and life expectancy. This is an example that I borrowed from Heather Cross, who is the statistician and um, that chart, it shows up positive association. The larger or the bigger is the cigarette consumption per capita, country by country, the higher the life expectancy of those countries is, right? So if you’re a cigarette smoker and you don’t read the… you don’t think about the graphic carefully, you may describe the content of that chart, the more we smoke, the longer we live, right? And that’s how, that’s what the chart is showing. What the chart is showing is that there is a positive association between cigarette consumption or life expectancy and vice versa, but that doesn’t mean that one of these is connected to the other in any sense. There’s a problem with correlation causation, there’s problem with ecological fallacies, there is a problem with Simpson’s paradoxes and many other things. So readers need to make an effort in sort of stick to the idea that a chart shows only what it shows and nothing else. Most of the other inferences that we made out of charts are inferences that happen in our brain. They are not in the chart itself and that’s a perfect example of that. Now, in a case like that, obviously, there’s always, there’s also a huge responsibility on the part of the designer. If I were to make that chart myself and publish it, I would add a very big caption warning people about not to see that in the chart. Right? This chart is not showing that the more that you smoke, the longer you will live. It just shows that countries that are richer on average can buy more cigarettes and countries that are richer tend to also have better healthcare systems. And as a consequence of that and many other factors such as, you know, diets and exercise and things like that, people also tend to live longer.
JS: Yeah, it’s really interesting because I think when people see any visual stimuli, they’re, they’re led to make conclusions or see patterns. Right?
JS: So, so this, this point of trying to say, hey, don’t, don’t draw a causal link between the two, um, try to only see, you know, try to only see the correlation, seems like a, I guess it’s a, it’s kind of a heavy lift, right, for, for designers.
AC: It is a heavy lift. But as I said, I mean I think that a designer can make an effort to add, you know, if you are going to publish a chart like that, there may be a good reason that you want to publish a chart like that. You may want to make a point about that association for some reason. You know, you can use some space in the chart to warn people about what the chart is not showing or what, what are the possible wrong inferences that you can extract from the chart.
JS: Yeah. It’s also a question of audience, right? Like where do you draw the line of what do I need to explain to what audience member, right. Like a scatterplot in an economics peer review journal isn’t going to need a lot of the explanation.
JS: But on the Washington Post website, it probably does. And so then you have these audiences in between.
JS: Like how do you think about it? I mean, you’re a journalist. You, you think about different types of audiences. So how do you think about targeting different audiences and trying to meet their expertise where they are?
AC: I prefer to put the emphasis on explanations. So, um, I tried to put myself in a frame of mine in which I assume that people know a little bit less than I believe they do. Right. So I try to add more explanations rather than less explanations just to avoid these kinds of problems. Um, the problem with that obviously that you can end up having visualizations that are a little bit over burdened with explanations, and texts, and annotations, et cetera, et cetera. But I’d prefer it that way. I think that again, as I said, you know, visualization can sometimes lead you to see patterns that are not really there or, or making inferences that are not warranted by the, by the data or by the, by the graphic. And I think that it is worth it to, um, warn people about that. Right. So there is a responsibility on our end to the…
JS: Yeah, definitely. Let me, I want to switch gears a little bit and ask about your process. I mean, you’ve written, let’s see, this is your, what fourth book? You’ve got a couple more in the works. I think we can talk about those. Um, this one is interesting because I know you’ve, you, this is a topic of interest you’ve had for a while and then, uh, was it last year maybe you did your visual trumpery tour where you sort of, you know, went around the world that looked like and, and talked about these topics. And I’m curious about your writing process and also how that tour and your conversations and your presentations affected what you wrote, how you wrote things that people said to you, you know, what, what was that experience like?
AC: Yes. The, the tour in the, I mean, between 2016 and 2018, I did a couple of visual trumpery talks also in 2019, but it mostly, they, mostly all of them took place between 2016 and 2018. Basically, the talks, these series of presentations, they led to the book, right? So I first put the slides together, gathered tons of examples that I had in my computer, added some more, et cetera. And I structured the talk as I talk, explaining the systematic ways in which either charts are designed to lie in purpose or the ways in which, in which we mislead ourselves or lie to ourselves with charts. Um, and I use the talk sort of unconsciously as a way to test ideas, examples, see how people reacted to those examples, notice what people understood or didn’t understand in the examples that I was, um, that I was presenting. And that shaped the, the book, because originally I was planning to do a book just about line charts. This has, these are just charts that lie. These are chart that lies for this reason, for another reason. These are misleading charts for this reason, for another reason. But what I realized is that that’s not what people need because if you only do that, you’re not giving people the tools to take advantage of charts. Because the, the title, the title is How Charts Lie is a provocative title, but the subtitle gives you a clue of what the book is truly about because the book is not a book about here, here’s a tons of, a ton of line charts or here are a ton of, a ton of misleading graphics. The book is a manual about how to become a better chart reader. So I added a whole chapter that is basically sort of a Grammar of Graphics light. The famous book, the Grammar of graphics. You know, how our graphic is a structure. How visualization is a structure? What is visual encoding? I explained to the general public what visual encoding is, right? So I’ve involved, you know, tons of pages to basically explain that how charts are red, right? The same way that we need to teach people how to read words. We can also teach people how to read visuals, how to read visualization. So the tone of the book is positive. Originally, the tone was a slightly negative, right? This is not, this is bullshit. This is a bad chart. This is whatever. And there’s certainly something loud, something about that in the book itself. Um, there are plenty of examples that are really, really bad, but most of the book has a very positive tone. It’s not, you know, it is not, the book doesn’t just say charts mislead us very often. It also says, but charts can be used to make us smarter, to make us a better, better human beings and more informed. And this is how to [indiscernible 00:22:08] I have to do it.
JS: I want to, I want to make you king of the world for a moment or at least king of the education system? Um, what would you change in the way people learn how to read charts from kindergarten all the way through, uh, through college? Like how would you change the curriculum?
AC: Well, you cannot really, I don’t think that we can really detach, um, graphical literacy or graphicacy from numerical literacy, also called numeracy. There’s a famous book title Innumeracy by John Allen Paulos, which is fantastic. It’s a fantastic book. I think that both things go hand in hand. We need to help people become more numerate, become more used to dealing with numbers or reason based on, based on numbers and, and then we’re going to also teach people, help people become more visually literate, more graphicate. Right? Those things go hand in hand. Now how to do it? I have no idea. I mean, I don’t know. I’m not an educator, but I’m not used to teaching small children. The way that, the way that perhaps we could do it in math classes is to spend a little bit less time making children, you know, do complex calculations by hand and spend more time discussing how the numbers that they see every day in the classroom apply or relate to their, to their lives. Uh, perhaps, perhaps use examples that speak to them. So more examples about music or movies or things like that. And then talk about, you know, how to reason about the numbers, about the, related to those topics that they care about, right? What is the album or the song that has sold more copies in the past? What is the song that has made more money in the past 10 years? Right? And you can use that to explain, I don’t know, adjusting for inflation, right? A song that was published this year obviously will make much more money than a song that was published 20 years ago, but the, that’s just an effect of, of inflation rate. If you’d done adjustment for inflation, then it will appear that way. So you can use that as example, as example, as an entry point to explain a complex idea or a complex issue. But again, this is just a very general idea. I don’t know, I just think that I, I’m more fond about the classes that sort of expand your mind by helping you see the multiple angles in which you can approach a topic rather, rather than just teaching people how to make comically complete operations by hand, which I also believe is necessary. It is necessary to multiply, but after you have done that 10 times, just [indiscernible 00:24:48] calculator.
JS: Well, what’s interesting about the, about the field of data visualization, right, is that it brings a lot of these different skill sets together. You’ve got the math and you’ve got the literature and you’ve got design and art and um, you’ve got even, you know, computer science. It’s bringing all these different skill sets and philosophies together into one area.
AC: Yeah. And not only quantitative fields, it also brings together, you know, rhetoric and journalism and narrative and the storytelling. It’s like, it’s a bunch of stuff, right? Yup.
JS: Yeah. So would you change the way people are taught visuals at the university level?
AC: Yeah, absolutely. So, um, actually this may inform, um, this idea may inform, um, one of the books that I have planned for the near future. Um, it’s a still a little bit vague in my mind. Um, but I would like to follow the path of Andy Kirk. You know that Andy wrote his book, um, with the idea that visualization is a process, right? It’s not…
AC: Yes. Just go deeper into that idea and write a book that talks about how to reason about visualization, how to make good decisions about visualization, not by applying cookie cutter rules, right? Which is how visualization is usually taught. Here’s a bar chart. Here’s a bar chart for these. Here’s a scatterplot. Here is a scatterplot for these and go deeper into the reasoning behind all of those rules. And that way I think that people will understand better when the rule is applicable and when the rule needs to be broken or when the rule can be basically just avoided or how to create new rules and how to expand the vocabulary of data visualization. So how, how to think about visualization, how to reason about visualization or how visualization designers currently think, right. That will be another way in which, um, in which people can learn. So I think that that’s the way to teach visualization at the moment to anybody who wants to learn it.
JS: Do you think the, the data viz field is, is pivoting in that direction in terms of what people are speaking about and writing about on blogs and on and on websites?
AC: Um, people who have been in the field for, for a relatively long time, absolutely. They are pivoting in that direction. Yeah. The field is pivoting in that direction. Um, but I don’t care that much about the people who have a lot of experience, right. They are autonomous in their own, right. I’m more worried about other people who are entering the field at this moment, right. We need, I think to find the balance between saying, you know, there are certain, um, rules, quotation mark in there, there are certain principles, there are certain conventions, there is that tradition in data visualization and you need to respect all that because there is a reason why all these things exist. But at the same time, it is also important to understand where all these conventions, principles and rules come from, which one of them are more or less supported by either evidence or logic or practice, etc., um, learn how they were developed, etc., and then learn how to break them or how to expand them or how to create new ones in the future, right. So, yeah, we are pivoting in that direction, but I think that we need to pivot perhaps a little bit more.
JS: Hmm. Interesting. I have the reading copy here in front of me and uh, I’ve been going through it again, I think this is like the third time I’ve read it. It’s, uh, it’s great. I’m really enjoying it and, um, I look forward to seeing it come out and, and make its ways around the world and see how, see what people say about it.
AC: Thank you, Jon. You’re very, very kind.
JS: Well, thanks Alberto. Always fun to chat with you.
JS: And thanks to everyone for tuning into this week’s episode. I hope you enjoyed it. I hope you’ll check out Alberto’s new book, How Charts Lie. Uh, it is coming out any day now. Um, and if you’re interested in seeing Alberto speak, uh, in person, he’ll be at the Urban Institute in October, uh, to talk about his book. Um, so, uh, stay tuned for information on that. That’ll be coming out in a little while. Um, and if you’d like to support this podcast, please check out my Patrion page or just share the show with, uh, your friends and your colleagues and review the show on your favourite podcast provider. So until next time, this has been the PolicyViz Podcast. Thanks so much for listening.