Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
A History of Data Visualization and Graphic Communication A History of Data Visualization and Graphic Communication with Michael Friendly & Howard Wainer
464 Plays3 years ago
Michael Friendly is a Fellow of the American Statistical Association, a Professor of Psychology, founding Chair of the graduate program in Quantitative Methods at York University, and an Associate Coordinator with the Statistical Consulting Service. He received his doctorate in Psychology from Princeton University, specializing in Psychometrics and Cognitive Psychology.
In addition to his research interests in psychology, Professor Friendly has broad experience in data analysis, statistics, and computer applications. He is the author of Discrete Data Analysis with R: Visualization and Modeling Techniques four Categorical and Count Data. He is also the author of SAS for Statistical Graphics, 1st Edition and Visualizing Categorical Data, both published by SAS Institute, and an Associate Editor of the Journal of Computational and Graphical Statistics and Statistical Science His recent work includes the further development of graphical methods for categorical data and multivariate linear models, as well as work on the history of data visualization.
Howard Wainer is an independent statistician and author with experience in educational testing and data visualization. He received his PhD from Princeton University in 1968. He has taught at The University of Chicago, Princeton University and the Wharton School of the University of Pennsylvania. He was employed by the Educational Testing Service from 1980 until 2001 and was the Distinguished Research Scientist at the National Board of Medical Examiners from 2001 until 2016. He is a fellow of the American Statistical Association and American Educational Research Association.
Episode Notes
Michael Friendly and Howard Wainer, A History of Data Visualization & Graphic Communication
Michael Friendly GitHub | https://friendly.github.io/HistDataVis/
Milestones Project: https://datavis.ca/milestones/
Michael Friendly Site | https://www.datavis.ca/
John W. Tukey, Exploratory Data Analysis
Sandra Rendgen, The Minard System: The Complete Graphics of Charles-Joseph Minard
Brit Rusert, Silas Munro, W. E. B. Du Bois’s Data Portraits: Visualizing Black America
Leland Wilkinson, The Grammar of Graphics
Isabel Wilkerson, The Warmth of Other Suns: The Epic Story of America’s Great Migration
Recommended
Transcript
Introduction of Guests and Book
00:00:12
Speaker
Welcome back to the policy of his podcast. I'm your host John Schwabish on this week's episode of the show. I'm very excited to have Michael Friendly and Howard Wehner join me. Michael and Howard are legends in the field and they have a new book on the history of data visualization. Really interesting book, especially if you've seen a lot of the discussion on Twitter about historical data visualizations, they give you a more thorough treatment than you're going to get on Twitter. So we talk about how they work together. We talk about their favorite visualizations.
00:00:42
Speaker
We talk about their favorite eras of historical data visualization. Really fun conversation. I hope you'll enjoy it. So take a listen to this week's episode of the Policy Vis podcast with Michael Friendly and Howard Wehner. Michael Friendly, Howard Wehner. Good afternoon. Thank you so much for coming on the show. Great to see you. How are you both? I'm great. As well as could be expected.
00:01:08
Speaker
Thanks so much
Exploration of the 'Golden Age of Statistical Graphics'
00:01:09
Speaker
for coming on the show. I'm very excited about your book, History of Data Visualization and Graphic Communication. If folks haven't seen it, they will check it out by the end of this particular podcast episode, which is going to be great. I've got a whole set of questions for you. And I thought we would start by talking about the book itself and what your favorite parts are. The book really starts
00:01:29
Speaker
at the beginning of time of how, you know, people were drawing on cave walls and using sticks to communicate data and understand the world around them all the way to the modern era. And I think we'll start with maybe, Michael, with you. Like, which era did you find the most interesting? What was the most enthralling for you to both research and write about? Thanks for having us, John. Let me start by saying this is sort of like the scene in The Crown.
00:01:55
Speaker
where Prince Philip asked the Queen, which is her favorite child? It's unanswerable, but I love all of these ages. But let me say, first of all, that the parsing of history into themes that have a particular order is one of the most important features of our book. Yet I have to confess, I do have a favorite child, and my sympathies are most with the golden age
00:02:24
Speaker
of statistical graphics. In a sense, this was the culmination of what had been building for over 200 years. The development of wide sources of data go back to the early 1800s. Theory of measurement and statistics, that goes back to the early 1600s with mapping the heavens and navigation at sea. Technology, the ability to reproduce
00:02:54
Speaker
exquisite graphs in full color was something that came into fore in the period from 1850 to 1900. In this period, there's also some of my favorite heroes in this history. Charles Joseph Minard, Andre Nichelle Gary, John Snow, Florence Nightingale, Francis Galton. Each of these contributed something that was new
00:03:22
Speaker
and magical in the representation of facts, important statistical and scientific discoveries. But the main thing that I love most about the Golden Age is this incredibly impressive collection of the album De Statiste Grafique, published in France from 1877 to 1899. The most incredibly exquisite sampler
00:03:50
Speaker
of all known graphic forms and inventing new graphic forms as they went along. Their topics were sort of mundane. How is our trade in wine and cotton doing with the rest of Europe? Where should we build a railroad? But they used exquisite graphs to show things in a way that captured the eye and captured the imagination.
00:04:18
Speaker
I spent five years trying to track down the complete set of the album, De Statice Grafique, with Howard and a bunch of other colleagues. We purchased the entire set. We owned them each individually, but now David Rumsey has acquired a new set and made them all available on his website in full high resolution.
Organizing the Book by Themes and Periods
00:04:42
Speaker
So when you think about organizing this long history, you mentioned the distinction between chronology and theme. And I wonder if you could talk about that a little bit more for folks who may not be as familiar with the history of data visualization stretching all the way back.
00:05:00
Speaker
OK, let me go on that. So when I was first cataloging and organizing the material in my milestones project, it struck me that each of the periods of time, 1600s, 1700s, roughly in centuries, had a coherent theme that went across not only the kinds of graphs that were produced or maps that were produced, but in terms of the important problems
00:05:30
Speaker
of the age. So the 1600s, I call the period of measurement and theory. This was the time that European countries were competing for markets, for discoveries of new territory, for what would become their colonies and sources of great wealth. Well, this took the combination of scientific measurement
00:05:52
Speaker
and recording of information about past voyages. Edmund Haley made an incredible isogon map of declination of the magnetic compass at sea. Like if your compass is drifting, you're going to get lost really soon. He did this from remarkably few observations, but that was an important contribution both to navigation
00:06:20
Speaker
and to graphical display using interpolated curves on a map for very little data. Yeah. So you mentioned a couple of times the interplay between statistics and graphical methods. And I'm wondering, and I'll turn this over to Howard, I'm wondering whether you think either could have existed without the other.
The Connection Between Data and Meaning
00:06:45
Speaker
No.
00:06:47
Speaker
I want more. I mean, maybe. Maybe there's no more to say, but maybe some more. Have you ever testified in court? The instructions you get from a lawyer is that the answers are yes, no, I don't know, and blue. Why blue? What is blue? If the question was what color car was they driving, the answer is not a blue 57 Chevy. It's just blue.
00:07:17
Speaker
The story begins in the contrast between Plato and Aristotle because it's the idea of rationalism versus empiricism. And empiricism, without the idea of empiricism,
00:07:38
Speaker
There's no need to have data. If evidence doesn't matter, then data doesn't matter. And if there's no data, there's no graphics. And so you couldn't have graphics unless you had a belief that data mattered. And in particular, if you wanted to make claims, you needed evidence to support those claims.
00:07:59
Speaker
Unfortunately, when Aristotle was making his point about evidence, he had Alexander the Great backing him up. And so he didn't screw around with Aristotle, because he had to deal with Alex. But it died out. At the end of the golden age of Greece, that died out, and it didn't come back again until Bacon in the 15th century, and then the other Bacon a little later.
00:08:22
Speaker
But really, empiricism didn't really get rolling until the British empiricists, in particular Hume. And it was only when the idea that if you wanted to learn something, you needed to have data, you had to have evidence, and you couldn't just guess. And that's not fully
00:08:38
Speaker
absorbed now either. And you have lots of arguments being made in the absence of data or in conflict of data. And that's going on and on now, to this day. But somewhere around the 18th century, where the British empiricists got rolling, this is certainly only in the West, people started gathering more data. And whether it was health data or crime data, or weather data,
00:09:04
Speaker
And somewhere along the line, they discovered that the best way to see what was going on is literally to see what's going on. And that brings in one of the heroes that Michael didn't mention, that's William Playfair.
00:09:19
Speaker
Uh an 18th century scoundrel, but he was the one that invented most graphs, right? Uh, you know, of course there were maps and things like that uh, but that's that's really the everything hangs on The philosophical basis of empiricism without that you haven't got anything right so you started with plato and aristotl So we're zipping forward now to to the last say 30 40 years and i'm curious um because these two things are so
00:09:49
Speaker
intertwined. Why do you think folks in math statistically data dense fields like economics and maybe sociology and statistics and mathematics, why are there visualizations just like often so bad?
Modern Visualization Challenges and Lessons
00:10:09
Speaker
Because
00:10:11
Speaker
Did I respond to that? You know, instruction can help cure ignorance, but not stupidity. And consequently, people are drawn to what's flashy rather than what works. And to determine what works requires work.
00:10:35
Speaker
running little shoebox experiments, and you ask not which graph do you like better, but which way is the wind blowing, and where did the storm come from? Michael, you're chomping at the bit. You go. Well, let me put it in a wider context. I think of statistics and the development of statistical theory as the glue that binds data
00:11:05
Speaker
to discovery and persuasion. So the very idea of taking the average, this was revolutionary in days when different observers were recording the transit of Venus and using that to calculate the shape of the earth. Well, you had three or four different people making the same,
00:11:32
Speaker
measurement, but they differed. The idea that you could take the average and that would be a meaningful thing rather than taking the one most trusted or the observer who was most trusted by the royal astronomer. What Graf's did in this history of combining statistics with data and reasoning
00:12:01
Speaker
Is it allowed people to see the patterns, the trends? What stood out as an anomaly? It's this confluence of data, statistics, theory, and graphics that I think is so compelling. Yeah. Now, Howard, you mentioned a few moments ago
00:12:27
Speaker
that people see things that look kind of showy. They think it looks kind of neat. And so that's what they create. I want to just read one of my favorite sentences from the book. So in the last chapter, you both write, visual displays of empirical information are too often thought to be just compact summaries that at their best can clarify a muddled situation. This is partially true as far as it goes, but it omits the magic. And I wanted to get your sense of the magic. Do you get a sense that
00:12:56
Speaker
You know, there are some folks in the field that are very dogmatic about it. You never ever, ever, ever make a pie chart. You never use anything that's circular. There should be bar charts and line charts and, you know, there's lots of phrases and terms that we could reference and use, but do you feel that that removes some of the magic from the field? Well, the magic comes from what data it is that you're showing. There was some guy who mimicked Minar's plot of Napoleon's march.
00:13:25
Speaker
one of the most wonderful plots that's ever been made. And he had data about AOL stock prices that he was showing. It wasn't all that interesting. If you don't have interesting data, you don't have interesting graphs. Now, we omitted what I consider to be one of the most important points, and Michael alluded to it. And that was in 1962, when John Tukey published an article called The Future of Data Analysis.
00:13:54
Speaker
In it, he pointed out that the role of data analysis and statistics is discovering things, trying to find things. He concluded that the best way to find what you weren't expecting is through the use of graphic displays. There were a couple of
00:14:16
Speaker
19th century economists named Farquhar, two brothers, they said that trying to get information from a table is like extracting sunlight, sunbeams from a cucumber. You know, it's in there, but you can't see it. And the ability to be able to see things that you hadn't expected was what makes the graphics powerful. And Tukey, being a world-class hotshot, when he came out in favor of this in 1962, it made it okay
00:14:45
Speaker
For the rest of us. So even by 1962, or maybe especially by 1962, Tukey was the person that everybody looked to. For certain kinds of things, yeah. For certain kinds, right, right, right. To say that this is okay, that this is the sort of thing that we can be doing. That's right. Yeah, yeah, interesting.
00:15:04
Speaker
Um, so I want to talk about historical data visualizations kind of more broadly. Um, I wonder what you would say to folks, maybe, um, Michael, you can start on this one. I wonder what you would say to folks who say, yeah, historical data visualization is neat. I've seen Menard. I've seen Nightingale. I've seen John Snow. They're neat, but there's not much to really learn there because
00:15:28
Speaker
I'm, you know, I have to do mine. I'm coding in JavaScript or I'm, you know, doing something for a mobile phone. And so, you know, they're neat to look at. They're kind of cool, but like there's really nothing to learn there. And I'm curious what you would say to a person who maybe feels that way.
00:15:43
Speaker
John, before I answer that, let me just go back to Menard for a second. So why was it that Charles Joseph Menard's incredible depiction of the near destruction of Napoleon's army was so powerful? It was essentially meant
00:16:03
Speaker
as an anti-war statement. He was appalled by the tragic loss of nearly all of the French army, and he wanted to point it out in a way that spoke to the hearts and minds of his viewers.
00:16:20
Speaker
E.J. Marry, who was the first one who noticed Menard's graph, said it defies the pen of the historian in its brutal eloquence. So that is part of the magic of graphs. On to the idea of, well, we've got all this software today, and why do we need to worry about it? Well, what we've learned in the most recent history of data visualization
00:16:48
Speaker
is that we really needed a coherent overarching theory of graphics and the production of graphics. And that came with Lee Wilkinson's Grammar of Graphics. I was pleased to note that you did, I think, the last podcast interview with Lee before he tragically died. So this provides an overview, an overarching framework. And most importantly, it creates
00:17:18
Speaker
an easy path between having an idea in your head, oh, I want to make a graph of COVID outbreaks and how it's been moderated by the introduction of vaccines or other things. Having that idea and going to something you can see on a screen or on paper is, I think, the most important contribution of modern statistical software.
00:17:48
Speaker
What that leaves out, though, is the question of graphical impact. Tukey famously talked about the idea of interocularity of a graph. The interocular traumatic test is that a graph, a good graph, should hit you between the eyes. You should know its meaning and impact immediately. Well, software doesn't help with that.
00:18:17
Speaker
What is important for a graphic designer or a graphic communicator is to have a crystal clear idea of the message they want to convey and think of, okay, I have all these tools, how can I use them to create
00:18:39
Speaker
a graph or graphical display that is impactful, that gets to the hearts and minds of my audience. Just like Florence Nightingale was successful in her radial diagram showing the deaths in the Crimea, she could have used a simple line graph.
00:19:00
Speaker
But that would not have gotten the attention of the members of the British parliament who were tasked with trying to see what they could do about the disastrous loss of life, not on the battlefield, but to people who got sick from septicemia and died from septicemia.
Enhancing Narratives with Historical Techniques
00:19:20
Speaker
We tried to get to this a little bit in chapter 10 to try to show how you could take a modern question, in particular, the question of the movement of African Americans from the south to the north. And Du Bois was desperate to show that. And he tried very hard to show all sorts of things. And if he had just borrowed some of the techniques that Minar had developed,
00:19:46
Speaker
he could have come up with something else. And so we made up this story of Minar meeting with Du Bois in Paris and over cocktails working out this picture. There was a song that Du Bois sang to Minar to get him to come along and help.
00:20:05
Speaker
I don't know if I can, if I play that, would you be able to hear it? I'm not sure. Why don't we give it a shot? And if not, I'll stick it in the recording later. Why don't you two talk while I try to find the goddamn song? Okay. Okay, so I was gonna ask, I was gonna ask...
00:20:19
Speaker
The title of the book is History of Data Visualization and Graphic Communication. I'm not really sure what the question is, but maybe the question is, how do you distinguish between those two things, data visualization and graphic communication? Or maybe the better question is, why do you distinguish between them in the title? Well, one of the things that struck me in writing the book was how this entire history or part of it was the rise of visual thinking, the ability of
00:20:49
Speaker
not only scientists or economists or statesmen like Playfair to think visually but for their audiences to think visually. So graphic communication is the idea of being able to tell your story in the way that will
00:21:11
Speaker
most resonate with the audience. So one story, an early story of Michael Florent van Landgren, the first statistical graph in 1642, he had the idea for a new way to determine longitude at sea. He wanted to sell it to the King of Spain. He wanted a patronage appointment. What he did was he gathered all the previous determinations
00:21:40
Speaker
and he wanted to show the king, oh my God, everybody is making extravagant errors. The results are all over the map. He could have presented this stuff in a table, but only a graph had the power to show exactly what he was trying to show that everybody else was totally wrong. And therefore you should fund me and give me a life appointment. Right, right.
00:22:08
Speaker
Interesting. So, Howard, do you have our song? I do. Because in the Paris Exposition of 1900, they had the first talking movies, and so they were able to record this song that Du Bois used to lure Minard to help him out in this. Here's what we were able to recover.
00:22:36
Speaker
There's a new decision, who felt they were laughing? There's a new thing, I'm wanting to see beyond control. Me now is it, my life.
00:23:09
Speaker
As you might've noticed, that tune was co-opted later on for a James Bond movie. I think we all data visualization folks try to strive to be like James Bond. That's it. Explain that again. That's Menard and Du Bois. Du Bois, that's right. Du Bois had data, and the reason he had data was that census started collecting data on African-Americans.
00:23:41
Speaker
I guess it was starting in 1870 for the first census. Up until then, they had collected data, but it was as property, like three cows and two slaves and that sort of thing. But starting in 1870, they were collecting data. And so with data, you'll be able to find out things, you can answer things. And Du Bois was wonderful at being able to look at the data and try to tell a story.
00:24:08
Speaker
And he chose graphs as a way of doing it. He had 56 different graphs that he displayed in Paris on the ways of black folks. But the big story of the idea of the migration, the great migration, was very hard to be told. And Minar, of course, provided the way to do that.
00:24:31
Speaker
Right. And so that's the story that's told in Chapter 10. And there's a Yiddish term that goes for that story. It's called a bubmamaitza. Do you know what that? I do. I do. My mom, who may be listening to this episode, will really enjoy the fact that that may be her first occurrence of Yiddish on the show.
00:24:52
Speaker
You know, I was thinking of Yiddish the other day because the Eskimos have 56 words for snow. And Jews have schmuck, schmekl, schmendry, so on. We've got 86 words for loser.
00:25:09
Speaker
There must be something relevant there. Something matched up there, yeah.
Creation and Future of the Book
00:25:14
Speaker
So I want to ask one last question on the development of the book. Maybe this is better as the first question, but I wanted to ask you, you two have known each other and worked together for a long time. I wanted to ask, you know, how did this book come together and what was the process like? So maybe Howard, you can start on the history here. The origin story of the book, as it were.
00:25:33
Speaker
Well, Macklin and I had been working independently on the general topic of data display for the better part of 40 years. And each of us had gone in their own direction. But it became clear, at least to me,
00:25:49
Speaker
that time was running out and we needed a coherent statement, that the field needed a coherent statement because it had been more than 80 years since there was one. And that was some guy's master's thesis. It was really good, but I mean, it's Funkhouser.
00:26:07
Speaker
And by putting everything together, I felt that we could not only lend some coherence to the field, but also highlight what are the areas that we don't know about. And by we, I mean Michael and me, or don't have time for. And so I felt that if we got this thing going, it would be a good thing. And Michael's done such beautiful work.
00:26:33
Speaker
The various bits and pieces that he's done over the years are really wonderful. And I know he was saving them up for a book that we're going to be ready in 50, 60 more years.
00:26:48
Speaker
And so it was mostly me saying, come on, let's get this done. And him saying, it's not ready yet. We got more to do. We got more to learn. And so we finally did get it done. And he's still working on the second edition because there's so much more we didn't do. Right. I'm not sure I have the wit or the energy for a second edition.
00:27:15
Speaker
I'm glad the first edition's out and it really showcases a lot of Michael's contributions over the years. Yeah, that's great. So Michael, what for you
00:27:28
Speaker
I mean, in any book, you have to sift and winnow down to get it done. Are there things in your head, big topics, big themes, big visualizations that didn't make it into the first volume? Oh, my God. In zikry. About half of the material that we originally had planned on,
00:27:54
Speaker
could not find a place in the book or went on the cutting room floor of editing. Our contract called for 50,000 words. Who knew when you're starting to write a book? What's 50,000 words? By the time we were finished, we were up to 150,000 words. So we had to fight tooth and nail with our editor at Harvard University Press.
00:28:20
Speaker
to compromise at 100,000 words. Nonetheless, there are so many areas that we did not explore. We omit nearly all mention of modern data visualization. Lee Wilkinson, who wrote a wonderful review of the book,
00:28:42
Speaker
privately said to me, oh, you didn't mention grammar of graphics. And I said, oh, my God, I am so sorry. But we decided to sort of basically cut things off around 1975 or so. Howard mentioned earlier in our discussion the whole idea of non-European contributions to the history, the rich history of data visualization and information visualization
00:29:13
Speaker
We don't have enough material, nor is there a coherent structure for thinking about those really brilliant non-European developments. But that is something that I think of, oh, maybe someone else will write.
00:29:32
Speaker
Well, let's hope so. I mean, this is a great book, great synthesis of the field. I hope everyone will check it out. The history of data visualization and graphic communication.
Episode Conclusion and Further Resources
00:29:41
Speaker
Michael Howard, thank you so much for coming on the show. It has been great chatting with you both. Well, thanks very much for asking us. Thank you so much, John. I really enjoyed your other podcasts and I'm looking forward to hearing this one. Thanks. Great. Thanks again.
00:29:56
Speaker
Thanks everyone for tuning into this week's episode of the show. I hope you enjoyed that conversation with Michael and Howard, and I hope you'll check out their book. I've linked to it and many other things in the show notes on the website for this episode.
00:30:10
Speaker
If you would like to learn more about data visualization, check out my NoWino community. Instead of cluttering up your inbox with newsletters and other things, I'm using Winno to send you two or three text messages each and every week about data visualization strategies, big and small. So head over to winno.app.policyvis. That's W-I-N-N-O.app.policyvis to learn more. So until next time, this has been the Policy Vis Podcast. Thanks so much for listening.
00:30:39
Speaker
A number of people help bring you the PolicyViz podcast. Music is provided by the NRIs. Audio editing is provided by Ken Skaggs. Design and promotion is created with assistance from Sharon Satsuki-Ramirez. And each episode is transcribed by Jenny Transcription Services. If you'd like to help support the podcast, please share it and review it on iTunes, Stitcher, Spotify, YouTube, or wherever you get your podcasts.
00:31:01
Speaker
The PolicyViz podcast is ad-free and supported by listeners. If you'd like to help support the show financially, please visit our PayPal page or our Patreon page at patreon.com slash PolicyViz.