Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode #140: Scott Berinato
194 Plays6 years ago
Scott Berinato is is the author of Good Charts: The HBR Guide to Making Smarter, More Persuasive Data Visualizations. Even though he’s a writer, he’s also a self-described “dataviz geek” who loves the challenge of finding visual solutions to communications and...
The post Episode #140: Scott Berinato appeared first on PolicyViz.
Recommended
Transcript
Introduction to Scott Berinato and 'Good Charts'
00:00:11
Speaker
Hi, everyone. Welcome back to the Policy Vis podcast. I'm your host, John Schwabisch. On this week's episode, I'm excited to have Scott Barinato join me. Scott is the senior editor at the Harvard Business Review. He's also author of the book, Good Charts, the HBR guide to making smarter, more persuasive data visualizations. And he also has a new book coming out that's sort of the workbook version of Good Charts.
00:00:34
Speaker
It's a very interesting conversation talking about all things data visualization related and the process by which he thinks that we should create good visualizations.
Supporting the Show
00:00:44
Speaker
Also, before we get into the show, I'd just like to ask you to consider becoming a supporter of the show. I have a Patreon page set up. The money through Patreon goes to help buy better audio equipment, better sound editing services and transcription services. If you've been watching the show or listening to the show for the last
00:01:02
Speaker
a couple of months, you'll notice that I've been adding transcriptions of each of the episodes for folks who may need them or want them for their own work. So I hope you'll enjoy this episode. I hope you're having a great end of the year. Happy holidays to you and yours. And so on to the show.
Scott's Role and Editorial Focus at HBR
00:01:23
Speaker
Scott, hey, how are you? Welcome to the show. Thanks a lot, John. Happy to be here. This is exciting for me.
00:01:28
Speaker
Oh, great. Well, I'm glad we're able to finally get this chat together. I'm excited to talk about the book and all of your recommendations, and I want to hear about this new project you have that's coming out, I think, shortly. But let's start by having you maybe talk a little bit about you, yourself, your background, and how you ended up as an editor over at HBR.
00:01:48
Speaker
Sure. Yeah. Thanks. So yeah, I'm a senior editor over here, which means I do a little bit of everything and I'm a bit of a free agent. You know, a lot of people specialize in the web or the magazine, but I do a little bit for every platform. Editors at HBR are a little different than other places in that we do a lot of author management. You know, we have academic authors who have great ideas, but maybe aren't the best writers.
00:02:12
Speaker
or can't really put their thoughts together in a way for a lay audience. And so we really help them do that. And I do that for a lot of topics, not just visualization and data science, but also I like to specialize in neuroscience and some other sort of what I would call more visceral topics than strategy and the softer leadership, those softer things. I like the real sort of hardcore scientific ones. I do a lot on office space. How do you coordinate office space? I think it's all part of my sort of visual approach to thinking.
00:02:41
Speaker
that the topics i'm attracted to have a real spatial and visual sort of aspect to them so it's a lot of time editing writing thinking about new ways to express stories for the audience which in media is all we ever think about now.
00:02:58
Speaker
And it's a lot of fun and I get to do a lot of chart making as part of this because I'm an editor who's known here for sort of making the designer's lives miserable by always trying to add a visual element or add some kind of visual storytelling to my articles, which sort of breaks their templates and things, but I think it's really important.
Journey from Journalism to HBR
00:03:16
Speaker
Now, are you an editor by training, by background?
00:03:19
Speaker
Yeah, so I started out as a journalist, middle school journalism in Northwestern after undergraduate Wisconsin, dove right into the tech journalism scene at a magazine newspaper called PC week at the time.
00:03:31
Speaker
Um, right at the front end of the dot com boom, which was fun. Um, and I got to cover a lot of, a lot of the craziness at that time. I got to cover the Microsoft trial, uh, so on and so forth. And then I sort of got into feature writing, um, really started writing more features about the people in technology, security, cybersecurity, things like that. Um, and at some point I decided I wanted to see how HBR, which maybe is sort of the.
00:03:58
Speaker
oldest of old media brands, right? It's almost 100 years old, how they were thinking about the future of media. And I just saw it as a great opportunity to be part of a transformation of a brand like that. And that's a joint here about 10 years ago, and I've been working on that ever since. Nice. Now you said a couple of things I want to make sure I catch before we move on. So first off, are you a badger? Is that what you said?
00:04:21
Speaker
I am. Go Badgers. Yeah. Yeah. I am. I am. We're both Badgers. No. It's like the best episode ever. It's great. I was just back there. I was back for the opening game this year, taking my daughter to visit the campus, and I was hoping she'd love it, but I don't think she's going to go there alas. Oh, that's too bad. I keep pushing. My kids are several years away from that, but I keep gently just planting the seed of Wisconsin is the place to go. Yeah, we had a lot of fun.
Tools and Workflow for Visualizations at HBR
00:04:51
Speaker
The other thing you mentioned was on the chart making and I'm curious what the process is like for you and for HBR. Is there, you know, do you have a certain suite of tools that are used for the actual publication, both online and print, or is it whatever works for you personally? You just make something that, you know, works for the, for the publication.
00:05:10
Speaker
Yeah, for a while it was really the Wild West. It was whatever I could get my hands on and when I could do it and I'd present them with something reasonable and then the design team would pull it into Illustrator and make it beautiful. And that was our workflow for a long time, not very efficient, effective in sort of the kind of hacking it kind of mentality, but we knew we had to get better. We now have a full-time information designer who works with us, who's great.
00:05:34
Speaker
And we use a combination of tools really depending on the types of visualizations we need to do, but we also have some very strong templates in Illustrator, especially for print, where we're pulling data in from tools like Plotly, Tableau, others, just exporting the SVGs and then designing them up into those templates into our host style in Illustrator, which seems sort of old school, but it's actually still the best way.
00:06:01
Speaker
It's interesting. I'm sure you've come across this, but the tools question is always the question I get first, right? When I'm speaking and what tools do you use? And the truth is there's no one tool that does everything well and the tools that do more things well, require more training and.
00:06:14
Speaker
And it's, I wish I could just say use X tool and that will solve all your problems, but it's not. So we do a little bit of everything. We do some high charts. I like to dabble in D3. I'm not a programmer by any means, but I do love using D3. I love the concept and how powerful it is. I use Plotly online a lot. So I use a lot of different tools so long as I can export SVG. And then we sort of design it up in Illustrator.
Challenges in Data Visualization at HBR
00:06:41
Speaker
Right. Now what about the authors? I mean, the HBR has a suite of authors. So are they creating things or are they pretty much saying, here's the data, here's the screenshot from my journal article that looks terrible. You go make something that looks good.
00:06:59
Speaker
Generally, it's the latter. I mean, they usually have good ideas, but really, and none of them I think would be surprised at this sort of academic execution is what I'd call it. I don't like to do sort of like good or bad. I'm always about, like people are trying, but you know, some things just don't reach a general audience the way they could. So some of the academic charts are really challenging, but they always have good data. So we always have good starting point.
00:07:25
Speaker
And a lot of it is us taking their data and then having a conversation with them about what it is they're really trying to show.
00:07:33
Speaker
taking it away, doing something to it, showing to them, and then refining with them. There is a lot of that. Now, I will say on the conceptual visualizations, which are pretty common in HBR2, your virtuous cycles, your two by two matrices, those kinds of things, they tend to fall in love with their conceptual diagrams, even if they're not particularly effective. And that's always a harder conversation to have. And I'm sure you've come across this. Some of the conceptual diagrams out there are
00:08:01
Speaker
are strange and bizarre. Yes. The funnel diagrams and the 3D funnels. The pyramids that are actually the 3D pyramids with camels walking in front of them and things like that. We just saw one. It was a 3D pyramid with a city skyline behind it. We couldn't figure out what the city skyline represented. It turns out it was just decoration. Of course. We get a lot of that.
00:08:23
Speaker
But you have authors who are at least I would I would guess they're sort of a selected group, right? Like they are writing for for HBR to reach a wider audience. So I would guess that their mindset is, let's make this as appealing to a broader audience as possible. Is that right?
00:08:39
Speaker
Usually, yeah, I think usually sometimes some authors are so enamored of their data and so enamored of showing all of the data they collect that they have a hard time sort of editing, you know, they just want to show everything all at once. And they think this is, I have to show everything and we've come across this, you know, a few times where authors just insist and we have to have these sort of really
00:09:02
Speaker
hard conversations with them about how what they're trying to show is actually not coming through. People are just seeing this sort of mess of bars or lines, and it's actually not effective in the way they think it's effective. They understand it because they collected the data, they've thought about the data, they know the data, but the audience is never going to understand that. So every once in a while,
00:09:22
Speaker
We have those hard conversations, but for the most part, I think the authors really welcome the opportunity to have this sort of transformation happen. I just had this experience with an author from Northwestern who was doing some really interesting stuff with network communications and network diagrams, sort of force-directed graphs.
00:09:42
Speaker
And he really welcomed my input and we really worked back and forth. And we turned something that came out of his statistical package that looked reasonably difficult to parse into something really simple and clear and understandable. And he was just thrilled with the end results. So that's usually the way it happens.
00:10:00
Speaker
Right.
Framework for Quick Chart Creation
00:10:01
Speaker
So that process leads, of course, to your book that came out, I think, a couple of years ago. Good charts. So I would guess that the lessons you talk about in the book are the ones that you are sort of applying to the authors that you are working with on a day to day basis.
00:10:17
Speaker
Absolutely. And I use the process. I actually lay out a framework in the book for a way to get better at doing charts in a couple of hours. The whole point of the book, the whole idea behind the book at the time was that we needed a guide for people
00:10:32
Speaker
who were intimidated, who felt sort of put off by what I would call the actual judginess of the charting community, who felt sort of the oppressiveness of rules and understanding when you can never do this or always do that or colors to use. And we're so stressed about the grammar of charts that they weren't getting anywhere. They just decided, you know what? I'm just going to click the button in Excel and paste it into my presentation. That's good enough because this world is hard. So we were trying to put our arm around the reader.
00:11:01
Speaker
And we have a framework in there that says, if you just do these steps in a couple of hours, you'll have something better than where you started. And I still use that process, and I'm still sometimes surprised at how well it works. But the whole point of the book is not so much to say, always do this, never do this. These are the rules. It's to say the right chart always depends on the context. And so we take a design thinking approach. First, you have to ask yourself, what am I trying to say? Who am I going to say to? And where am I going to say it? If you don't do that before you make your chart, you're sort of dead in the water to begin with.
00:11:31
Speaker
And once you understand that and feel that and understand why that matters, I think people start to get to better output pretty quickly.
00:11:38
Speaker
Yeah. So what do you think is the biggest challenge for the sort of folks that you work with who have a lot of data? I mean, it sounds like they working with data, they're able to make charts. What is the challenge? What's the roadblock that keeps them from taking the chart that they make in Stata or SaaS or, or whatever it is and making it something that you would be happy with publishing on the HBR platforms?
Team Approach in Data Visualization
00:12:03
Speaker
I would say two things. The first is time. It's always time, right? They just need more time to do this. And this does take a little bit of time. I can't lie and say, oh, you can make it happen just as fast as you're doing it now and have better output. That's not the case. But I think also, and I'm writing a feature for HBR about this right now, which will come out in January. I think something's happened where the notion that the people wrangling the data and parsing the data and analyzing the data are the same people who should be visualizing the data.
00:12:31
Speaker
And I always advocate, I call it the unicorn problem because these companies invested billions in getting data science operations set up and they have invested very little on what I call the last mile, which is the communication part of it. They find all these insights and then they can't tell the bosses what they're finding because they can't communicate well. And so I always say a team approach is much better and that you have to sort of
00:12:57
Speaker
amass different talents for different tasks and put those together on teams. And it's not a throw the stuff over the wall and let the designer fix it kind of thing. It's literally sitting together and working as a team. And that this team-based approach was lost, I would say, in the early 80s. And in fact, I have some old books. I don't know if you're familiar with them. There's one from 1912 by Willard Brinton called Graphic Methods for Presenting Facts.
00:13:21
Speaker
You know that book by any chance? Yeah. It's fantastic, right? And in that book, it's just sort of a, well, here's your team and you have your draftsmen and you have your data entry people and all of this. And then even in the 50s and 60s, this woman who worked in the US government, Mary Eleanor Spear,
00:13:37
Speaker
She was a great chart maker and a great advocate for visualization. She in her book said, well, here's your team. These are the three people on it. And the idea was it was just assumed this was a team effort. Right. But then then Excel happened and Chart Wizard happened. Right.
00:13:54
Speaker
And you get Chart Wizard and it's like, wait, I can just click on a button and I get a chart and I can make it 3D and put a glass box around it. And this is incredible. And that kind of convenience trumps the notion that, you know, we can have better output if we think a little bit about what we're trying to show and get a few people together to put that out there. And I think we're getting to the point now where it's so important to communicate visually again, that that's no longer good enough and that we have to start putting these teams back together.
00:14:21
Speaker
Yeah, I mean, you mentioned the tools earlier, and there's no tool that does everything. I actually, you know, don't want there to be that tool that does everything, because I want there to be these specialization of skills where you've got, you know, the graphic designer does what the graphic designer does really well. And the
00:14:39
Speaker
and the data person does what the data person does really well. And these groups sort of overlap, at least in their appreciation of what these other groups could do. But one, like the Lord of the Rings tool rule them all makes me a little bit nervous. Yeah, I think that's right. I think I agree with you. And I wish it was a little more organized world. It's hard to, when you have no experience with it, to get the experience to know which tools to go to, when.
00:15:05
Speaker
for sure, but also I think it's a lot of times I mentioned this team approach and like you were saying, you know, there's the graphic design tools and the statistical tools. It's a lot of throwing things over the wall, right? The data science team will come up with something and they'll output it in ggplot or Python, whatever. And they'll throw it over the walls to say designers, you know, make this pretty, which is not the right approach at all. And I really also advocate for designers taking some stats classes and some stats folks taking a couple of design classes.
00:15:35
Speaker
I'm not saying become an expert in these fields, but at least develop a fundamental appreciation for what they're doing and how they're doing it, why they're doing it. And I've seen that when designers take stats classes especially, they come to understand some of the challenges that they have to tackle, like uncertainty and other concepts that we're dealing with here. And then when the statistical folks take a design class, then
00:15:58
Speaker
They can sort of let go of their fear that simple is bad or simple is hiding things or simple is manipulative, which I think is oftentimes their fear that something that's pretty and beautiful and clear and simple is not holistic or objective. I mean, I love also for data and stats people to understand that design is not making things pretty. It's more than that. It's moving the content forward and it's functional.
00:16:23
Speaker
Absolutely. And especially when you get to the sort of the storytelling sort of aspect of in the presentation aspect, design is actually fixing your audience where you want it fixed. Right. It's actually it's actually moving their attention at your will. And it's it's a it's a really impressive thing if it's done well. It's it has nothing very little to do with what I have a design friend who's just absolutely brilliant. And she calls it she calls it the difference between designing and styling. Right. Styling is making it pretty and giving it the right colors and all of that. But design is actually
00:16:52
Speaker
you know, sort of fixing the experience. Right. You mentioned rules earlier. Yeah. And I wanted to ask, are there any data visualization rules that you think are absolutes in in the
Rules and Logic in Data Visualization
00:17:07
Speaker
practice? It's funny. Yes. Because somebody just asked me this the other day and I was trying to think of one. And the only one I can think that you really can't break ever, ever, ever is to make categorical data continuous. You know,
00:17:19
Speaker
where you have a line chart connecting things that should be bars or something i mean that's just wrong right it doesn't make any sense it's illogical i don't have a lot of rules where i'd say never ever do that i'm not a pie chart vigilance you know i've had arguments with people i
00:17:36
Speaker
I have a friend of mine who published a chart with a truncated y-axis, and somebody tweeted to him he thought it was a thought crime, what he had done. I thought, that's a bit extreme. I don't think that's true. The rules I like people to think about are the ones where they're based in visual perception and neuroscience.
00:18:00
Speaker
There are things that you do to the brain when you showed a visual that if you upset the conventions and the heuristics in our brain, you actually make it harder to understand, right? A simple example would be if time goes right to left on your chart, right? It just doesn't make sense to our brains and we have to stop and think and anytime
00:18:18
Speaker
We're sort of flouting conventions that way. We're making it harder for people to see the information or the idea clearly. There are times when you know somebody is going to spend time with something that it's okay. But in general, if time is going right to left, that's really hard for us. Or some other conventions like that, sort of heuristical conventions that happen in our brain. If you did a world map and you put it, I'm putting finger quotes up now, upside down, right? So the south pole was at the top.
00:18:47
Speaker
There's absolutely nothing wrong with that. In space, north is not up. But our brains are so used to seeing that that you just make it really hard for people to process the information in that visual. So those are the kinds of rules I think about, but I would never say never do a pie chart. I would never say never use blue for women or pink for men or anything like that.
00:19:14
Speaker
I always say, if you draw me a sketch and it says exactly what it needs to, I'd rather have that than the most professionally designed chart that has the wrong information in it. Yeah. So when either you're writing yourself or you're working with authors there or others, what's the thing that gets you most excited about communicating data or data visualization, either one?
00:19:37
Speaker
Uh, two things. I think, um, I always look for what, uh, is called, well, some folks call the bliss point. I really am always in seeking the bliss point. Um, and I'll define that for you because it's not a term that's sort of real, but a guy, I know Kirk Goldsberry. Do you know Kirk Goldsberry? Yeah. He's a, he was at the Kennedy school. He's done basketball viz for a long time. He now works for the San Antonio Spurs. He's a data scientist, a great guy, smart guy, but he calls it bliss points. When you create a visualization,
00:20:06
Speaker
that is so clear and so directly applicable to the audience that they see it and there's almost sometimes an audible gasp. There's this moment of understanding and then what settles in after that moment of understanding is it almost feels like it was always true and I always knew it.
00:20:23
Speaker
And when you create charts that just have that instantaneous effect, but then can be used for sort of deeper diving. I'm always excited by the chance to create a chart that people just get and understand and affects their thinking right away, that kind of thing. I think that's one thing. The other is, especially we're working with academics who have very heady ideas, very complex ideas, helping people clarify their thinking visually, even if it's not something that's published.
00:20:52
Speaker
But if we can just sit down and actually sort of maybe draw a map of their ideas or something, I really love to actually work through challenges visually. I love the idea that even if it's something that's just going to sit in my sketch pile, it's something that actually helped me get to the solution I needed. And I do this with my stories all the time when I'm writing.
00:21:11
Speaker
I'll map them out, I'll sketch them, I'll draw shapes that represent where I'm going with the story. And sometimes they serve as great guides and the world will never see them but they helped me get to a point. I'm a real advocate for sketching all of the time and trying to just come up with visual solutions to problems or even using visuals to solve problems. So those are the two things that sort of excite me the
Visuals in Public Discourse
00:21:32
Speaker
most. And then anytime
00:21:34
Speaker
You know, Twitter is a weird and terrible and awful and wonderful place, but anytime public conversations and debates can happen with visuals, I think if it's being done responsibly and thoughtfully, I think it's a really powerful way to have good public discourse. I think it's
00:21:54
Speaker
not only necessary, but almost important. It's really important. These problems we're trying to solve in the world today are so complex, and many of them are so difficult to really understand with a critical eye that the only way to do it is through visuals. I always look forward to really smart debates based on visuals.
00:22:18
Speaker
Right. So let me, I just want to reflect back then on this, the complexity versus the simplicity, a graph that says, Oh, I, you know, I get that right away. You know, the, the example of time going right to left is sort of.
00:22:33
Speaker
you know, throws us off a little bit, maybe, but then there's this balance between sometimes we want to have a graph where people say, Oh, I get that right away. You know, that line is going up over time. And other times we want people to sort of get into the visualization, explore it a little bit, maybe take some more time. So how do you think about that, that balance, I always come back to there's this sort of famous
00:22:53
Speaker
graphic, Iraq's bloody toll that, you know, people sort of show, right, which is the deaths in Iraq and the vertical axis is positive deaths, but it's going down or sort of the, you know, right to your point about the time. So, but that one, you know, sort of people get in and they explore it. So I wonder how you think about those two, I guess, competing thoughts when it comes to creating a visual for people.
00:23:15
Speaker
Yeah, and this goes back to that idea of context I talk about in the book all the time over and over again, right? So let's start a presentation. You have what, 400 milliseconds or something to get somebody's attention. And the biggest mistake I see in presentations is they will present the Iraq's bloody toll chart and they'll spend most of their presentation explaining how the chart works rather than, you know, the ideas in it.
00:23:36
Speaker
And so that's a case where simplicity rules above all for me, right? But then, Iraq's bloody toll in a newspaper where I'm sitting down and it's eyes on paper or even eyes on a screen, right? And I'm having this sort of intimate experience. Yes, I want more detail, more depth, more things to explore, more surprises as I look through it or even play with it and change the variables. I think anytime I know my audience is expert,
00:24:04
Speaker
I want to provide them more depth, more detail. If I'm putting a chart on Twitter to the world, that's your lowest common denominator. The more complex, the less it's actually going to be processed except for by people who really want to engage with it, which may be fine.
00:24:22
Speaker
But I sort of had these rules of thumb about what is the expertise of the audience? What is the experience? Is it 50 people looking at a screen and me? Is it one or two sets of eyes on a piece of paper or on a screen? And I sort of use my sort of intuition about how much detail I can provide then. And even then, I really love it if a chart can
00:24:44
Speaker
both give that initial impression and then allow you to explore it further. I think those are the sort of the most successful charts where you look at it and you get it, but then you want to spend time with it. Yeah, definitely. Definitely.
Interactive Design of 'Good Charts Workbook'
00:24:57
Speaker
I want to give you a chance to talk about your new project as like a follow-up workbook to good charts. That's coming out in any day now? January. Oh, January. Okay. So we have some time to prepare ourselves. Go ahead and pre-order. That's fine by me. Okay.
00:25:12
Speaker
Good charts workbook is the follow up and it really comes out of the doing lectures and doing workshops with folks and oftentimes they'd be very inspired in this. OK, but how do I do this when you're gone? How do I start? And that question sort of prompted the entire workbook. How do I start? And so this is a chance for you to really build your skills by yourself. And what we do is we set it up sort of like a crossword puzzle book where we give you a challenge. We'll show you a chart that may be in
00:25:39
Speaker
good shape and maybe in bad shape, we'll just show you a dataset and we'll give you three or four things to do with it. And then a little bit of blank space though, you're going to need extra paper if you scribble like me and sketch like me, I'm just all over the map. But a little bit of sketch space and then we have a discussion. It's not really an answer key because
00:25:57
Speaker
I find the idea that I could give you the right answer, obviously not right because I'm all about context and I don't know your context. I'm giving you some context. I'm giving you how I solved the problem, but I encourage people to disagree with me and say, I don't like how you solved it and we'll actually give you an email address where you can send me your solutions to the challenges.
00:26:17
Speaker
And the challenges in the book are really, you know, there's sort of two levels of them. The first are just to build those basic skills, sort of like a workout, right? So there's a chapter on color and there's a chapter on clarity and there's a chapter on choosing chart types. And these are like your, you know, lifting weights and all of that. And then in the back of the book, we give you the massive challenge where you put all that together and build a presentation based on some data we give you.
00:26:40
Speaker
I'm really hoping people love it. I had a lot of fun writing it actually. And I know the people who are working on putting it together now, they've had a lot of fun doing the challenges as they're trying to put the book together too. So we're pretty excited about it.
00:26:51
Speaker
That's great. Well, I'll link to it. I saw it on Amazon when we were talking earlier, so it's out there, folks, and preorder it. And of course, I'll put links to all the things we talked about, including your existing book that people, I think, should definitely pick out. Scott, thanks so much for coming on the show. This has been a really great conversation, really interesting stuff. Thanks a lot. Love you, John. Thanks so much, and I love your site.
00:27:14
Speaker
Oh, thanks a lot. And well, thanks to everyone for tuning into this week's episode. I hope you're enjoying the every other week format this year. It's giving me a little bit of a breather, which is nice. If you have comments or questions or thoughts for Scott, please get in touch with comments on the site or on Twitter. And I'll put links to everything that we've talked about in the show notes. So you can check out Britain's book in the Iraq bloody toll. And of course, Scott's work over at HBR.
00:27:39
Speaker
Thanks for tuning in. Until next time, this has been the Policy Vis Podcast. Thanks so much for listening.