On this, the 21st episode of The PolicyViz Podcast, I am very excited to welcome Edward Tufte to the show. As you might expect, I was excited to talk with Professor Tufte, so this episode is quite a bit longer than the usual episode. We talk about his art and sculpture, data art, the state of data visualization tools, and the future of data communication.

I’d love to hear from you, the listener, about this episode. Do you have thoughts on Tufte’s views of data art or data visualization tools? Did you take his workshop? If so, what was your experience?

And as always, should you have comments about the show or suggestions for guests, please let me know using the comment box below or on Twitter. And please rate the show on iTunes–your ratings help!

Transcript

Jon Schwabish: Welcome back to the PolicyViz podcast. I am your host, Jon Schwabish. I’m here with a very, very exciting guest on this week’s episode, I’m here with the one and only, Edward Tufte. Professor, welcome to the show.

Edward Tufte: Thank you.

JS: Thanks for taking some time out. You’re down here in Virginia, giving your famous workshop standing in the huge ballroom. So I’m very excited to talk to you. Let’s start with some of the work that you’re currently working on.

ET: There are three or four main projects. First, I’m doing my one-day course, and I do about 30 courses a year, and I still love it after all these years. We’re up now 275,000 people who’ve come to this over the years, and I love teaching. And the course is always kind of teaching my next book, they get all four books now, but I’m teaching more out of my current manuscript, and I’ve always done that, sort of in one book I have, and trying it out. It’s a good way to learn to talk through things to a class. You have to start to – you have to understand things.

JS: You have to [inaudible 00:01:24] right.

ET: You gain more understanding. So I’m trying to gain self-understanding. And so, I’m teaching the current book, documentary film, which is going so slowly that books and videos will probably be the same thing by the time [inaudible 00:01:42] it’s called the Thinking Eye. I regard the most important interface as the interface between the light comes into the eye and goes into the brain. So I don’t like the software metaphor, but the idea is to improve the software that how people think, and all my work has kind of been about that secret, how to make people smarter, but this is overtly now, how to think analytically, and judge evidence and reason about it, and skills to develop an attitude. It’s kind of on the background, it’s sort of an aspirational autobiography that I wish I could do all these things that I suggest that people should do with it. So I wish I could follow all my words. Some of the advice I try hard to do myself, it’s very hard.

JS: Yeah.

ET: So it’s sort of an aspirational. And to improve analytical thinking, first, about seeing and how to see better, and then about reasoning about what you see; and then, I think this is most important, producing, making something of the seeing. As Steve Jobs said, real artists ship. That is they make something and they execute [inaudible 00:03:20] do it.

JS: Yeah.

ET: And that’s how it has effects and becomes tested, and then there’s a cycle, after you go through this of editing, which again, is now trying to see your own work with fresh vacation eyes and, innocent eyes, and to think more, and to think afresh about, and then produce. And that’s very important that goes over and over, is that cycle.

JS: So is the book about the process through creating a data visualization or a graphic or is it sort of more general about creating [inaudible 00:03:59] producing a creative product?

ET: It can be teaching, it can be poetry, it can be writing, it could be a movie. It’s about anything that somebody else looks at. The [inaudible 00:04:15] is pretty narrow. This is about, and I don’t believe ever in pre-specifying a mode of production. In other words, a lot of people come and ask me at the course, I have to talk to the higher ups and they’re a little impatient or little stupid, how can I use visualization; and I say, do whatever it takes, don’t pre-specify, don’t even pre-specify a dataset, and don’t pre-specify for heaven’s sakes a method; you should use words and images and sock puppets and maps and charts, and that’s a very important point in the book is this spirit of whatever it takes. And that’s a very different spirit than being process oriented. A lot of scholarship is things like how to use data visualization to study economics. Well, the real question to me is, how do you answer that, how to answer some economic questions, and do whatever it takes for doing it. And that’s one of the big points of the book actually, which is it’s a little bit muddling through, but it breaks out of this [inaudible 00:05:25] where you use conventional methods versus pre-specified method. And there’s this big literature that’s always a little soft, using data visualization to study topic X, and they’re usually not a contribution to either field, that they don’t have quite enough content knowledge of either, and they try to gain credibility by combining these two things that people haven’t done before. That’s not a contribution, it’s making a finding about economics, or making a finding about data visualization. In fact, you bring a method to a new field is [inaudible 00:06:01] of course, any competent scholar may use different, but back to the thinking on.

And so, that’s going to have four essays, some on the idea about margins and edges, and frames, and surrounds, and what goes on in the interaction between the frame context, and the material inside. The thinking essay is the big thing, and there’ll be an essay about call scene around. This is about seeing in 3D space, and I’ve been learning about that very hard for 10 years, doing sculpture, and I’ve been showing a lot, I showed it at Fermilab and the oldest contemporary art museum, I had a gallery in New York for three years, and I’ve been doing something which is very hard for ours, especially sculptors like selling pieces, and I do it because, partly, I think I understand design in flatland, nobody ever will understand all of content in flatland, but I understand flatland. And you want a real interested problem, you see we’re good in flatland that reasoning about a cell on a spreadsheet or reading a poem – can that same kind of analytical intensity apply to as we see the world at focus and discipline, and simplicity of content which is not through the world. And so, everything is so luscious and so complex, and often so empirical of strategies and seeing in the real world.

So when I have a flat, that’s what I sculptures call a painting or engraving, the flat on the wall, and maybe two or three months, I don’t see it anymore, and I have to move it to see it with fresh eyes, every time I go out in the sculpture fields, it’s different, the light is different, I’m moving differently, the light is moving, the pace is moving, it’s from different angles, it’s raining, the dogs are running around, the season is different. And it is such a pure, wonderful, intense environment, contemplative and often beyond words, when people have a 234-acre sculpture farm, that will be left to put my foundation in perpetuity as open space, which shows the work of one artist, my work. And we have an annual open house or open one day for a year, each year, and our fifth one is on October 17, in Woodbury, Connecticut, this coming Saturday on October 17. And there are signs up in driveway, diamond signs, one of them that’s driving says shut up and look. It’s undiplomatic, but I mean it, that conversation uses probably half or two thirds of our brain processing power. Well, why not for a little while, devote all the brain processing power to see, it’s amazing the difference. Some of the land I have has old New England stonewalls, we were out walking along them, and I said to my friend, let’s stop talking; and after about 10 minutes, you could hear the famous noise far away and then the light seemed to change. And so, you could look underneath the trees, and the shadows were no longer blowing out the black; and the whites, light reflections off the snow were no longer blowing out the white. So it’s like a perfect photographic day where you have a gentle, filtered light, except it was happening in our brain, because we were all our brain power. And so, it increased the dynamic range of the eye, which is already pretty good, and so the blacks weren’t blowing out, and nor were the whites. And I just was so thrilled by that, the only thing difficult was the footsteps of my companion on the leaves became annoying.

JS: Conversation couldn’t cover that up, right.

ET: Right. And then also people change their experience when they don’t talk. So a couple will start to split up and walk around the pieces and be possessed by them; and when you walk by somebody, they won’t acknowledge you under this condition that I am talking.

JS: Yeah.

ET: And so, seeing graces have replaced social graces with that concentration, and this really came true – Paul Ekman who is a great psychologist of emotions and nonverbal behavior and expression of emotion, visited the sculpture grounds. He’s an old dear friend of mine, and he went out and he meditates and he’s done books with the Dalai Lama, and he said, they were beyond words. And I started crying, because he really got it. And I was so happy that somebody with his capacity…

JS: Background…

ET: His power could do that. So that’s the sculpture part. And a lot of things that I learned in flatland are more so in flatland, you have positive space, a figurine ground, and painting and diagrams and everything. Well, in the real world, you have the material, and you have airspace. But sculptors like me and even really super ones like Richard Serra agree or say that air is a material. And when I’m building something, there are great big pieces, they have 2 million pounds of stone [inaudible 00:12:22] big pieces, there’s more talk between my colleagues who are riding the backhoe in welding and so on, about air than about the material; and more talk about the interface between air the material, and also the joints between them rather than the body.

JS: Right.

ET: And so, it’s kind of eerie, it kind of changes the three dimensions, it’s not fixed the way it is on flatland. And if you walk around, as the light changes, and so it’s very empirical, you know you have to look at, there are certain strategies for looking that they are part of the thinking eye, so I’m looking at a three dimensional object, and what kind of [inaudible 00:13:04] and I look at folks, and then I turn my back and walk away, because you can’t walk away fast enough to see how the thing changes, you have to, because it changes gradually. Or if you walk around a piece, you have to walk fast. So I turn my back on them, and then I turn around suddenly, after I’ve gotten 10 meters away. And so, I see with fresh eyes, I don’t see this gradual shift. And so, that’s what I mean by a strategy for the thinking eyes, very practical in a way, about surprise your eye, keep your eye fresh, move at a pace where you can really detect change. So it’s not too smooth, you can detect before and after change. So all this – and it’s all hands-on, I love the physicality, it’s not like the computer screen. It used to be the way it was in design a long time ago where you cut and pasted and had X-Acto knives and all of that, and there was a craft hand, but that’s gone, probably for the good, but it’s gone.

JS: Have you investigated or looked at some of the data art that people are doing where there’s – unfortunately, I forget her name, but there’s a woman who makes baskets based on biological data – have you looked at any of that, have you gone to any of those exhibits to see what people are doing sort of blending those two?

ET: I will, looking at this as an artist only, it’s not first grade art. My view is that on all things, I only care about enormous excellence, I don’t want to see anything else. I fought the war against stupidity for a while, maybe about 5% is in fact the war against stupidity, against PowerPoint and [inaudible 00:14:57]. But I’m not really interested in that, it attracts people’s attention because it’s critical. But in my life, and almost all my work is very positive, and the whole basis is the identification of excellence, the careful study and celebration of excellence. And maybe the fourth time I’m looking at reading maybe being a little skeptical, but there’s so much to be learned from all the stuff in the past, and you see that in my books, which go back, they are in 20 countries, the best information in the universe is Galileo. He had terabytes of information through his telescope, the resolution, and his engravings and so on. And he just could see so well, and he can think better than anybody, and he’s doing science and nature. And whenever I come out of ideas, I look through his collected papers, and, oh my goodness, look at that.

JS: Yeah.

ET: So as an artist, that kind of work doesn’t count. That it’s like an interdisciplinary gamut kind of way, where it’s difficult in universities, usually interdisciplinary work doesn’t quite make it and either, I’ve done a little bit of data art, my Feynman diagrams which was stainless steel, and I showed them at Fermi labs, so I had real physicists looking at them, that was really fun.

JS: Yeah, that was interesting for that.

ET: And one of them said, pointed at one of my diagrams and pointed at where subatomic particles interacted, and said, how did that miracle happen. In real science, the word miracle is not a good word, not at all. And I had kind of half anticipated it, and I said, well, all these little fudges are particular constants in quantum mechanics. And so, this is just a little fudge. It’s a virtual on [inaudible 00:17:06] like a virtual spot. But that was so wonderful, because it was reassuring, like, the science was okay, but I saw them really as an artist.

JS: So you’ve mentioned all the books and sort of looking back, so I want to ask you to look back, you’ve been – the first book came out in 1986?

ET: ‘83.

JS: 1983. So you’ve been teaching these classes for 30 some odd years.

ET: Well, I taught them at Yale for quite a while, I took early retirement. I left Yale in ‘19, when I was 59, because the teaching was going so spectacularly well. My joke is that I retired from Yale in order to teach and do research, that I was unencumbered by it, and I didn’t have to worry about income because of the successes of the books and the tour. And so, I started teaching the courses in 1993, and I had two books then, I had Envisioning and then the first book. The first tour is called the kitchen tour, because we were [inaudible 00:18:17] in a kitchen. I want to name it, like the Rolling Stones…

JS: Yeah, right, it’s a very practical tour.

ET: This was [inaudible 00:18:26] but also, I wanted to get the word out, I didn’t want to sit around New Haven waiting for people to come because they’re not going to come. You’ve got to go out on the road…

JS: And you’re self-publishing, which at the time, I suspect, was a new sort of thing.

ET: Yes, and it’s a glorious story, but I didn’t want to react [inaudible 00:18:47]

JS: Okay.

ET: So that was the beginning, because I had two books, I used to go out with the books in the trunk of my car, make a little pile, and now we have movie vans and roadies, and 700 gigs later and 270,000 people later, it’s still, I really love it, I even almost like flying, because I’m going to a place I can do something I enjoy.

JS: Yeah. Okay, so 20 some odd years of teaching this particular workshop, I’m curious if you have seen a change in the types of people attend, the question you hear… And then let me broaden that question into what you’ve seen in the field of visualizing information and visualizing data, and how you’ve seen that sort of change over the last 30 years, and maybe where you see it heading over the next 30 years? So that’s a lot I know in one question.

ET: Right. Well, I can say about the people who come, I do zero audience research, I pay, I don’t think about the audience at all, I’m not interested in reactions. I talk to a lot of people, I have office hours, but I’m too busy writing books and selling books and going to the course to do market research. We don’t have time for it, because we’re too busy shipping stuff. So I think the history of data visualization has undergone a tremendous transformation due to one big thing, which is the resolution of our display devices, is now getting close to the resolution of paper, and is almost worthy of the human eye-brain system. So in years past, everybody was looking at a Dell laptop, and the resolution of that was 8% of the resolution of paper. You can now get a 4K screen, which is 4.5 times HD, for $800. It’s the high resolution screens that made the smartphones possible, the iPad possible, the 5K Apple which I just love possible, and the software hasn’t caught up at all. The graphics a lot of them are still doing second semester product things in programming, but it’s in a new language. There’s not, you know, the software is way behind, particularly on the windows side, way behind. And it’s the hardware that’s made visualization possible. I love it. [inaudible 00:21:20] so much because I wrote the books for paper, and most of all, the design work was done under fivefold magnification. There are jokes in those books that you can only see maybe three, four. And so, it’s like a dream for me, that now I could see on a backlit screen, back in this spectacular color, and interact with it, and enlarge it, and I’m a big fan of paper, but when I look at it on a good screen, it just glows…

JS: Yeah, it does, yeah.

ET: And art is just amazing, not quite like the painting, but in some ways it’s better than being there, because you can be alone with it and enlarge it and almost touch it. And a great piece you can’t normally [inaudible 00:22:10] that intimate with, so it’s the intimacy. And the software I think has been generally disappointing that I had a diagram that shows the – that’s originally from the New York Times, that shows the weather for every day for a year with temperatures and highs and lows and normals. And so, somebody sent me a thing, hey, I did this in R. Well, my students in 1984 were running something called Calcomp, and at Harvard Graphics there was an event, and doing those things. And so, what gain is there in a sense that it’s now being done in 20 different languages?

The other thing is Bret Victor said something very interesting the other day, which is that there’s a huge gap in the computer world between an idea and the implementation of it. And that hit home, I first made Sparklines in the mid-90s, I was consulting for Hewlett-Packard, and they were going to have a Unix box in every patient’s room, medical box, or maybe [inaudible 00:23:31] and they sent me protocol flow sheets in medicine, which was like a spreadsheet time across the top in categories, empirical time, not category, but empirical measure. And then what was what happened, the [inaudible 00:23:44] happened. And they’d be very long and very empty in [inaudible 00:23:48] spreadsheets. So I got these, and then in the margin, there was a little space, and I would accumulate all that history and make a little thing, a pencil thing, a little data word. And I was teaching my students that in the late 90s, well, so last week on the Apple Watch, I saw Sparklines, Medical Sparklines just right there, and made it to Epic which is a big medical data thing two years ago. That’s a long time for that to happen. It made it – actually it was interesting – it made it to Excel pretty early on. I was really – and they didn’t screw it up. Well, I was surprised they did a very good job. It made it to Google Analytics pretty early on. But the actual day to day implementation, finally, it’s been implemented quite a few places in sports particularly, but still to see it as a piece of brand new, a nice thing, like that, that’s a long time.

JS: Yeah.

ET: But Bret Victor was saying that’s common, the things that people have thought of years and years ago are finally coming through. There’s a lot of baggage and inertia, certainly in Microsoft, there’s a terrible baggage problem with all those people out there. Part of the baggage though is created by software houses that use damned proprietary formats, and it’s an awful problem. I’ve been working on some medical patient data. I did some papers long ago, and they’re now coming and – 94, partly because of recovery money, there was 20 to 30 billion in the recovery for medical records for electronic one, and that’s been a delight to the accountants. But you see medical electronic records all over the place, but you don’t see them in the survival data for patients. But I’ve been trying to push that a little. And the big problem is there are at least 1000 devices in hospital, each of which has his own damn proprietary format. It’s like a tragedy…

JS: And they don’t talk to each other.

ET: They don’t talk to each other at all. It’s so awful, I’ve been paying attention to what’s being done at Yale, and what’s being done at major at Cleveland Clinic. So the way that Yale communicates, say, cardiology, with Cleveland cardiology, is by something called fax. People under 40 don’t know what that is. You go into an active practice, and there’ll be 10 side inches in a day of faxed medical records, even though they both have the same electronic records. They can’t talk to each other.

JS: No, they can’t talk to each other.

ET: And it’s this problem that the programmers or the software houses think they own the content just because it happened to pass through their format. And the first person who said, accepted proprietary formats in the government and in universities and everywhere, that’s terrible, and they should write contracts. We don’t do proprietary. Period. We want this. Only bidders…

JS: Would open.

ET: [inaudible 00:27:13]

JS: Right.

ET: Because the costs now are enormous…

JS: At locking those things away.

ET: Yeah.

JS: Right. In the data visualization world, we’ve seen this sort of explosion in Tableau and Plotly. Do you think they’re all just sort of swimming on the top of the waters, not really diving into the deep stuff that’s really going to help make a difference?

ET: I think that it can make a great deal of difference by information throughput, and you can certainly see the incredible difference it’s made at the graphics news reporting at the New York Times, with scientists that had big data since Galileo, and they still – their work remains still at the cutting edge. The best visualizations in the world are found in the journal Nature, not at the spy agencies or visualization shops or supercomputer centers, with practicing scientists, whoever [inaudible 00:28:10]. But in terms of the quality of the credibility and quality of inference, they don’t make much difference. Because let’s say, there are 20 major threats to inference. And so, one of them is small sample size, only one, there’s still specification there, regression toward the mean, cherry-picking, and all the other things. And so, it can make some difference, particularly in how you communicate with people. But it hasn’t made too much difference in exploratory, yet, real scientists have solved the problem already, and they’re so far ahead of social science, except maybe for the times. But that’s where to look, and a lot of what the times has done is not visualization, it’s now being receptive to evidence other than words – words and photographs, that was what they did. They now do imaging of data, and they do a lot more data analysis. And so, that’s what’s made them, and the visualizations are terrific and access to the world, and they’re wonderfully done. But that’s only part of the revolution, it’s bringing numbers to journalism. That was always hard for them to do. I trained some journalists quite a bit at the Woodrow Wilson School years ago, and they were word people or they were photographic people. And the times really did beautifully by calling them graphics news reporters. It was about news, and there are 40 people in that department, only one of them is a designer, my student, Jonathan Corum, and the rest of them are statisticians or economists or reporters who have learned visualization, and some of the stories they write themselves. So it’s done to report, not to use the method, but to report the news.

JS: So do you think it’s about – so you’ve mentioned it’s an interdisciplinary work – in the new journalism, do you think it’s about individuals who have all these different skills or is it about the team?

ET: Well, I think it’s clearly about the team. You’ll notice that on the really big projects, there’re five names at the time, well, of course. But what’s interesting is that some of the people who do the graphics also do the reporting, so Jonathan Corum often reads the scientific articles and so on, and then turns them into visualizations by talking to the scientists and reading their work. And that integration between content and design has always been central to my work, I’ve [inaudible 00:31:10] essentially on the inside that it’s all about the content, not about the particular methods – I’ve done now for 30 years, I’m still shocked that people think that’s an amazing insight.

JS: I mean, I always use Ben Shneiderman’s quote, which said visualization’s about insight, not about pictures, and all rests on the data.

ET: Yeah. Well, it’s about explaining content, explaining things, yes, exactly.

JS: So where are we headed – so we’ve got all the new tools, maybe people are starting to build these teams like the Times and other agencies – where do you think we’re headed in the field of, I’m not going to call data visualization, I’ll say communicating data.

ET: Or reasoning about data.

JS: Or reasoning about data, right.

ET: Right. That’s thinking.

JS: Yeah or thinking about it…

ET: Yeah, thinking about it. My proposal, which I teach in my class, and which is in my website is called maps moving in time. So it’s 4K or 6K video, so no more sticks and ticks, 10-20-30 sticks and ticks and still land, but we should have the information throughput of video. I never did dare look at films for many years, from a data point of view, because they were so much information, it was scary, and they knew what they were doing, and they had infinite amount of money, it was frightening. But now, so the demo I have is of the Swiss mountain maps, which are a contour map, and then there’s a slow panning over them, and that slow panning leads to gentle 3D. Most 3D gives you a headache, or you have to wear funny glasses or sticks you in the eyes, it’s too aggressive. But the panning of the Swiss mountain maps, you can see a ski lift between two mountains with the peaks, it’s just incredible. And so, this has the information throughput and design power of a classic, probably one of the best maps ever done, there are contour lines and it’s just amazing, incredible typography. It’s the Swiss Alps’ great content too. But combining that with essentially infinite scroll, and probably not letting the user interact too much, because they’ll get impatient, it’s a slow pace, and so, they’re getting this tremendous information throughput on a 4K or 6K screen, and we’re talking 6000 by 4000, and there are just amazing things; and they’re getting that with intense design, great map and contour map with video throughput. And so, you’re cutting edge of resolution of video, and at the classic cutting edge of the layout. And so, I’ve been doing some things and there are demos on my website of maps moving in time, and so, it’s a complete shift. The problem is that the production requires tremendous amounts of data, and tremendous – well, not – the computational power is almost trivial now, but it requires lots of data, and it’s not easy editing video. It’s probably the square of the [inaudible 00:34:47] compared to still have something like that, or maybe it’s, yeah, you’re going for things, the area’s squared, and but movies are kind of almost volume. So it’s almost like the cube slightly. And so, see, again it’s aligned, it’s saying that the gain is in the hardware, video, well, software [inaudible 00:35:11] and then in the screens. And so, I don’t know if people are going to do their own data visualizations. But in real science and at the New York Times, and for serious big time presentations, that’s what – it’s already starting to happen. And that’s really powerful. So you have the patience and care with a gentle scan, but the rush of information with the video. And so, it doesn’t have the impatience of television, which is where they jump and stuff, that has that kind of concentrated analytical eye with the slow panning. And it’s the fundamental human act of thinking, which is in this small [inaudible 00:35:58] material you want to find the diamonds, the targets, the intervention points, the relevant points. And so, it’s aimed all information display is to support human cognitive skills. To me, the fundamental principle of design is that is to let the user perform the fundamental analytical tasks that people engage in when they reason about data. So understand causality, make comparisons and think in a multivariate way, [inaudible 00:36:33] and that’s the point, and that leads to designs.

JS: And do you think that visualizations can help people better understand the data, sort of going backwards a little bit, from, I’m going to present something to the user, and still help them understand sort of the whole network of how that visualization was created, I mean, that’s either teaching statistics through visualizations or teaching sort of complicated topics or big numbers or whatever it is, through visualization, do you think people can ultimately understand what’s under the hood?

ET: Well, it’s fairly high powered stuff, and I assume that just as places like Apple and Twitter have made many – consumerized many things that they will do it. And there might be some nice things, and there’ll be some people, probably young, bright young people who will do amazing things with accessibility to those tools. And so, they’re going to be much better at these things than we are. I just hired somebody who’s 24, and she’s out of this world, she’s a humanist and was studying Medieval, but she knows R. That’s scary. She knows LaTeX, she knows R, you know, wow. Okay, so I think as you think as students, young people, who are quick learners and so on, you’ve got to learn these things, and especially things like coding or things like that before age 25, it’s all over. I’ve tried to learn R several times, I can’t even hardly get past data entry, I get [inaudible 00:38:23] spreadsheet.

JS: Right.

ET: So, of course, it’ll somehow be – not commodified, that’s a terrible – and I hope it won’t be like R, that it will be accessible, it won’t be, you know, it will have a general thing, and there won’t be any interest in it, and the kind of…

JS: So something that’s more universal for even the lay user who can sort of go in and…

ET: Right. So if they can – okay, so they can make a reasonably luscious statistical graphic, some kind of fairly detailed map, and put their data on it, and then, okay, they pan over it. So, that’s now giving, essentially, it’s giving interaction, but it’s not letting the user completely set the pace. I think that’s very important that it’s hard enough to think and also to do interactions. Often that’s the kind of separation. The way I use most of my tools from R and from Illustrator and things like that, is I stand about three feet away, and I say, Elaine, move that a little bit to the right, I say to my design assistant. Because I find it hard to do a serious program at the level of Illustrator or R or InDesign, and think about that and get frustrated now and then, and have to solve a puzzle. And also to reason about the data, which I’m trying to do. And so, I think that now may be that some editing could be, I don’t know, it could be done by voice, and you can say do this or something, but you’re asking for a very hard combination, because these are big league serious programs, and they always have to have a fair amount of them, and there’s no real big solution. It’s like film editing, that the distance between a real film avid or final cut, and the difference between what we can do is a lot narrower now on our devices, but it’s not.

JS: It’s not simple.

ET: Yeah, it’s not simple, and it’s not real film, and it’s inherently complicated and hard. And so, clever interface won’t help much. I think what it means is, if you want to do that stuff, get a bright summer, be a summer intern.

JS: a young person who knows how to do the code.

ET: Exactly.

JS: Right. So you spend a lot of time, I would guess, looking at what people are creating, and the conversations that go on. When you see what people are talking about, when the creators are talking about what they’re making or people who are critiquing them, do you have a sense of how the community could do a better job of that? And maybe they’re not, maybe they’re doing a great job, maybe the conversation is great, but do you think there’s a community there where perhaps there’s changes to be made, and ways in which the whole field can sort of move forward?

ET: I think, if you look around, there are some really excellent critiques, and that’s just part of my usual strategy to find really great things and think about them. So the science article that took apart Google Flu is an amazing piece done by people at Harvard and MIT, and it makes excellent and obvious point that of the 20 threats to validity, little data [inaudible 00:42:17] is only one. And so, Google made errors naive, econometric errors, you know, model searching, and they had a very brittle model, which, of course, crumbled upon [inaudible 00:42:30] classic. And the Working with Time Series and stuff, the econometrics was naive. And the one thing I liked – the one side effect of the science article about Google Flu was that it did use the word hubris quite a bit. That’s what I mean by a really great critique. I don’t think – I think much better than critique is the style of my books, which is only 5% of the war against stupidity, which is, rather, the best critique is to make something wonderful to celebrate excellence, and say, this is our standard. So one of the very important things I teach in the classes is any diagram that you make, any table, any statistical graphic, any project management chart, anything that looks like any kind of visualization, you put it next on your screen to Google Maps, and you ask your IT department and your software people, how come I can’t put 120 words anywhere I want on my display, how come I can’t use light colors so I have five different layers, how come I have 40% computer administrative debris in my things, and they have personally none; how come I have to put, because of corporate guidelines, five logos on my thing, they have one little tiny thing, play in the big leagues? And so, that’s what I mean by this kind of comparison, and it’s not critiquing the thing, it’s say, here is a model of excellence. Usually when people are faced with a model of excellence, they say it’s too difficult, and the higher ups won’t understand it, and then, in response, you say, Google Maps is the most widely used image in human history, and you can – that’s refreshed every couple of days, it’s in every language, every country in the world, and people are using it, not just to look at, but to find out where they are and how to get there, and people use that information richness all the time. And so, that’s the kind of critique, I mean, occasionally, it’s nice to take something apart, especially if they are – Nate Silver does some nice critiques. And you’re usually taking apart the forces of evil. You should remember, when you do that, is that your allies cheat or do bad things just as much as your opponents, that your opponents are not uniquely and statistically are cheating; it’s just as true of your allies. And so, a lot of critiques are motivated by a disagreement with the substance, and it now comes a statistical question. But you got to remember, give me a critique of something you like philosophically, there is a role, it’s very active, of course, in peer review, that happens, and it’s also active is that the junk is never footnoted again. That’s the real sorting of quality, and I don’t want to look, I’ve got, you know, I probably have 10 or 15 years left, I don’t want to spend my time looking at mediocre stuff. There’s so much excellence left, and the chances are that looking at five pages of Galileo will be much better than looking at a week of Twitter or a month of Twitter or a year of Twitter. I enjoy Twitter, and it’s great fun, but it diversifies my point of view, and I get to see all this stuff quickly.

JS: Yeah.

ET: But I have a very long time horizon, both forward, and then I’m trying to do stuff that people are going to read for a long time, the challenge is to do forever now, and plus going back in history, and why should this day now be any better than some day, five years ago, or a day back – a Galileo day. And so, there is such a recency bias to our thinking, and to the web, and to, obviously, Facebook and to Twitter, it’s enormous. And everything is stacked by the most recent on top, so you don’t – I have my website, the threads are organized the other way. The oldest is on top and it grows as you go down. I think one of the very important things about the thinking eye is to think hard about your time horizon, and what time horizon you’re working for, but also from what time horizon. It took me until I was in my early 30s to realize that I did not have to read the quarterly journal such and such, because it would turn out, there was only one article out of 25 that would ever be cited again. And even the best article might be cited only a few times, and they wouldn’t even be cited by the author’s mother. So it’s to get out of that daily flow that many of us have to live in, and I’ve been really lucky to escape that, and to be able to have this long [inaudible 00:47:57] time horizon. And there’s so many wonderful things back then, or back, you know, last week, and they’re often more wonderful than today. And that was a great help, I’ve learned now is to think, look very long, far ahead, don’t do proper nouns, do verbs and do principles and do forever things, like some real scientists do with nature’s laws, do that. And then, the world of the past, really smart people have been doing visualization and explaining, and just spend a day in Galileo’s collected papers, and just look through the illustrations. They’re like 26 volumes, and I just do 10 pages, and I haven’t, you know, there’s another [inaudible 00:48:46]

JS: Yeah. Well, I think on that note of looking into the future, it can inspire us. I’d like to thank you so much for coming on the show.

ET: Okay. Terrific. Thank you.

JS: And thank you all for listening to this episode of the policy of this podcast. Until next time, thanks a lot. Bye-bye.