Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode #88: The Big Book of Dashboards
196 Plays8 years ago
Welcome back to the show! There are lots of decisions to be made when it comes to creating a data visualization. And when you pull multiple visualizations together into one view, the decisions compound. That’s why on this week’s show,...
The post Episode #88: The Big Book of Dashboards appeared first on PolicyViz.
Recommended
Transcript
Introduction to the Podcast
00:00:11
Speaker
Welcome back to the PolicyViz podcast. I'm your host, John Schwabisch. Welcome back to the show. On this week's episode, we are going to talk about a new book, The Big Book of Dashboards, Visualizing Your Data Using Real World Business Solutions. If you haven't checked it out, you certainly should. And I'm very happy to have two of the three authors with me today, Steve Wexler and Jeff Schaeffer. Andy Cockreave, the third author, is unavailable today, but I have
00:00:37
Speaker
I have the first two authors, which is really what matters, Steve and Jeff. Gentlemen, welcome to the show. Thank you. Great to be with you, John. How are you guys doing, okay? Absolutely. Delighted to be here, and we get to make fun of Andy, and he has no defense. That's right. He has no defense. He's stuck in meetings all day, so we'll take some hot shots at him while we go. At last, we've got him just where we want him.
The Journey of Writing 'The Big Book of Dashboards'
00:01:02
Speaker
Now, how long did it take the three of you to write this book?
00:01:05
Speaker
All told, it's been a little over two years. Probably about two years, because the whole gestation period of elephants were starting to feel sorry for us. It's about the right amount of time. So after two years, we can finally talk about Andy when he's not here. So let's start. I want to introduce each of you to the audience in case they're not familiar with your background and what you're all doing. And maybe you can each talk about yourself for just a minute or two, and then we'll talk about the book. So on today's show, we have Steve Wexler, who's the founder and principal of Data Revelations.
00:01:34
Speaker
data visualization consulting firm. And Jeff Shaper, who is the VP of IT and Analytics at Unifund. And Jeff and I have actually taught together in Baltimore, we've done a couple of workshops together, so we've known each other for a while. So, Steve, why don't we have you start. If you could talk a little bit about your background, how you got into data visualization, the sort of work you're doing now. And then Jeff, why don't you take a turn and then we'll dive right into this great new big book of dashboards.
Backgrounds of the Authors
00:01:59
Speaker
Sure. This is Steve.
00:02:00
Speaker
And as with most things in my career, I got into this accidentally. My background and training was as a musician, but I also had an interest in technology and computers. I ran a software company for around 20 years. Then I took my first full-time job working for an organization called the E-Learning Guild.
00:02:20
Speaker
They are an organization passionate about using technology to help people learn and perform better. And I came on board as their director of research and emerging technology. And I came in with the idea of, instead of printing static reports, why don't we make these interactive dashboards? And people could discover for themselves and filter the results of the survey questions. The organization had tens of thousands of users. So you could survey people and say, what works, what doesn't, what tools do you like, what don't you like, et cetera.
00:02:50
Speaker
and came in with this idea now 11 years ago to do these things. And they said, wow, that sounds great. What tool do you want to use? And I said, well, I don't know. Let's see what's out there. And then tested a bunch of tools and found one that I loved and still love. And that's Tableau. And kind of got into data visualization accidentally as part of this gig about 11 years ago. So like almost everybody coming at it accidentally,
00:03:20
Speaker
Jeff, what about you? What's your path to getting to the big book of dashboards? Well, Stephen, I have one thing in common there, which we were both musicians. I was a musician in a former life and moved into the business world. I've been in the financial services industry now about 22 years. So I've been doing that and really focused on IT and analytics for most of that time and was doing data visualization as part of the work here at my company.
00:03:49
Speaker
I guess I sort of fell into it accidentally as well. I was doing some presentations on some work of what our company was doing and the head of the department at the University of Cincinnati happened to be in an attendance and about six or seven years ago now asked me to build out a data visualization course for the university and been kind of doing that ever since. So I'm now adjunct at the University of Cincinnati teaching data visualization.
00:04:17
Speaker
And like Steve, I do a lot of Tableau work. And it's been a fun ride. Great. So new book, big book of dashboards.
Defining Dashboards Broadly
00:04:28
Speaker
I want to talk about a number of different things. But I want to start by defining a dashboard, which I'm going to guess was probably a lot of conversation amongst the three of you. But let me read the big definition here on page 14. Well, Roman numeral 14.
00:04:44
Speaker
Everybody knows that's right at the beginning of the book. So the definition here is a dashboard is a visual display of data used to monitor conditions and or facilitate understanding. Can you guys talk a little bit about that definition and how you view dashboards? Because I think that definition in some way can sort of encompass all visualizations and all visualization types. But I'm curious how you think about it as it applies to the idea of a dashboard.
00:05:09
Speaker
That's a great question. I'll jump in first there. It was a long topic, and way back before we were doing any writing in the book, I was exchanging some emails with Steven Few, and that was one of the things that he stressed to me in this adventure, is that we really need to think about what that dashboard is.
00:05:31
Speaker
that dashboard definition because as we move forward that's really you know sort of pivotal you know foundational i guess to what what you're going to have in the book and what you're going to call uh... dashboard and so uh... you know we did a lot of soul searching on that one we did a lot of definitional searching uh... for example you know steve's definition was uh... really focused on a single computer screen and monitored at a glance so we really cannot
00:05:58
Speaker
Talked about that. Does it have to be monitored as a glance? Does it have to be on a single computer screen? If the fact that it's printed out into a PDF, does that not make it a dashboard anymore? Which Stephen Few would say, once it's printed to a PDF, it's not a dashboard anymore. So we really talked a lot about that and thought about that. And so we intentionally came up with a broad definition, thinking that a dashboard
00:06:23
Speaker
to us could be any number of things that is broadly related to data visualization. I think there's some things in the data visualization field that go beyond that when we start talking about more advanced visualizations like network graphs and things that show that aren't necessarily used in a day-to-day business environment. But I think you'll get a sense, John, as you've seen some of the dashboards in the book, that it is a pretty broad definition.
00:06:51
Speaker
I'll add my two cents to that, which is we went through this whole process. We at one point defined what an interactive dashboard is, what a static dashboard is, and so much time was being spent on trying to define the thing. And we realized our focus is if you have this particular business predicament, this problem, this challenge, here is the collection, here's the chart or collection of charts that when logically assembled together will shine good light.
00:07:22
Speaker
on that particular business problem or scenario. And we didn't want to get into the whole issue of, well, that's a static dashboard. That's a dynamic dashboard. That's interactive. It was, you got this problem. Here is how other people have solved it and have solved it effectively. And this is why this collection works well and why other things that you might be tempted to try maybe don't work so well. Now, in your experience in working with people in the business community who are making dashboards,
Common Dashboard Types in Business
00:07:51
Speaker
What do you find is the most common type of dashboard? Is it the interactive dashboard and is it the type of interactive dashboard that is more of a presentation type thing or one in which me and my few colleagues sit around the computer and dive down? Or is it more of the static PDF type? There's a list here of the sort of things that might be considered dashboard and one of them is a PDF showing key measures.
00:08:16
Speaker
In your experience, where do people end up on this sort of spectrum, as it will, of different types of dashboards of interactive, to static, to things that allow you to explore, to things that are explaining results or analysis? In my world, it's all of the above that you mentioned, John. I would say, and I'm speaking in generalities, of course, but in my experience, I will never get the accounting department and attorneys off of paper.
00:08:45
Speaker
So it's just the nature of what they do. So when, when that comes up, things are often printed, their dashboards for compliance are often printed. And yet when we go into a monthly compliance meeting, we'll try to put it up on a screen. We'll interact with it. We'll have a discussion around it. We'll filter it by a state or by a certain, you know, complaint type. Um, and this is one of the dashboards that was featured in the book. We use that dashboard at a regular meeting on a monthly basis. And so there is interaction around it.
00:09:13
Speaker
but it can also be printed and handed to somebody and it works in that format as well and was intended to do that. So in my experience, we have dashboards that fit all of those definitions. Completely agree and runs the entire gamut and you may use the interactive dashboards to find where the most interesting things are and then you may copy and paste those things into a PowerPoint and share that with other people.
00:09:41
Speaker
One thing, Jonathan, that I got from your book is know your audience and in our case not just know your audience but meet with your audience and find out how they're going to use it, what they look at, how they interpret and hone what you're doing to meet those needs. Do you find that a lot of people forget their audience when they're making dashboards? That they're working with the data and they're making something for themselves to understand their data and then they bring it somewhere else and they have forgotten the audience or the person they're sharing with doesn't know anything about the topic?
Understanding the Audience for Dashboards
00:10:11
Speaker
Or perhaps it lines up a little better because the first thought is about exploration. You've got to know who's going to be using this thing and constantly be thinking, is this going to help my audience do what they need to do? Otherwise, you're building something for yourself and not for your audience. Someone asked, what's it like having two other authors on this thing? And did you get into big disagreements?
00:10:38
Speaker
And the debates and disagreements were, they were wonderful. They were enlightening. They were enthralling. But it wasn't like we were having artistic differences. We're not trying to make art here per se. Everything was about, am I making this thing as clear as it needs to be? So I miss having Jeff and Andy when I'm doing my practice, when I'm building things, when I'm trying to make interactive dashboards. I want the feedback. I want to know, is this thing as clear as I want it to be?
00:11:07
Speaker
And is it making sense to you? So as you went through, so there's a lot of, there's a huge collection of different dashboards in here and I want to, I want to talk about those in a moment, but were there certain ones where the three of you, uh, really disagreed on things where, you know, I know Andy Cockrove really loves 3d exploding pie charts. Um, now he's not here, so we're allowed to make fun of them. Um, so are there, were there certain dashboards or certain approaches that you all didn't see eye to eye on?
00:11:34
Speaker
I would say that, no, I think we had, it was interesting because any dashboard that was not created by us, we would set up a call with the designers of those dashboards and kind of go over it in detail. And then we would almost always have some suggestions. You'd see something and say, wow, this is a really great dashboard, but can we really put that in the book with this donut chart or, you know, that's kind of an extreme example.
00:12:00
Speaker
Maybe the color should be changed or maybe something is not as clear to us. And so there were more often than not minor adjustments made to some of those dashboards and we would all provide feedback into that. It wasn't Steve didn't jump on a call by himself and go down that process. We did it in collaboration and I think that's what Steve was alluding to that.
00:12:21
Speaker
We learned a lot from each other in doing that process from our own designs. We put ourselves through that. Nothing that I created was put in the book without them looking at it and critiquing it and vice versa. So it was really a great experience there. So talk a little bit about the types of dashboards that are in here and the balance between the dashboards that the three of you created versus the one that were created by outside folks.
00:12:48
Speaker
You know, how did you solicit those? You've already talked about sitting down with the author and talking about it. But talk a little bit about the collection of dashboards and where they came from and how you went through the process of evaluating them.
Collecting Dashboard Designs
00:12:59
Speaker
Jeff, I'll take this just because Jeff's kind of a hero of this story. Initially, my idea for this was, well, we're going to build all of them. We're going to come up with common scenarios that we've run into, and we're going to build all the dashboards. And I'm so glad we didn't go down that route. And it was Jeff who said,
00:13:16
Speaker
Why don't we get other people to submit their work? Because they're fantastic approaches and ideas that the three of us may not have considered. So for me, there were certain things that I'd just been seeing a lot that my clients and my students were having trouble with. And I wanted to make sure that we addressed it. An example is showing churn and turnover.
00:13:43
Speaker
in an organization with whether it's employee turnover for HR or subscribers and I've seen a lot of people struggling with that or people who were trying to automate the creation of a balanced scorecard type of thing and I just wanted to say oh my gosh if you're going to get into using a whole new tool and want to automate this why not make a much better
00:14:08
Speaker
collection of charts that that's better than that red, yellow, green mess that you were doing before. So of the 28 dashboards, yeah, a hunk of them were done by us. And then there, I don't know how many we got all told, probably somewhere between 60 and 70. Some of them just didn't make the cut, some of them were similar.
00:14:31
Speaker
So it didn't make sense to include it, and there were some great ones, and Jeff suffered some springs and arrows with this that he helped nurture along, and then the organization didn't get permission to include them in the book, and I was really sorry to see those missing there, but the real master stroke was, let's get the community
00:14:54
Speaker
to submit dashboards. And we very much wanted the community to be the widespread data visualization community. So we contacted people in the Excel world, the Dunn-This world, the Power BI world, the Spotfire world, the Click world, and obviously, and of course, the Tableau world. Can you talk a little bit about the tools that people are using to create dashboards? And when it comes to these different types of static versus interactive,
00:15:20
Speaker
Are people using different tools for each of those different types of visualizations, or is it I'm going to make an interactive visualization in Tableau, and then I'm going to save it out as a PDF, and it really doesn't matter in terms of whether it's going to be static or interactive.
Tools for Dashboard Creation
00:15:35
Speaker
I'm going to use this tool for both of those types of visualizations. I think that's important, and that even goes back to what Steve said in the definitional. When we were looking at an interactive dashboard versus a static dashboard,
00:15:47
Speaker
I think these tools have really changed that for us. That didn't used to be the case where you didn't have the ability to move data around like you do today where you can drop a drop down filter for state or for region or for segment or whatever it is and even do ad hoc on the fly in meetings. I've done this where the report is done by state and all of a sudden you want to see it by category type or something and you can make a quick adjustment to your visualization.
00:16:16
Speaker
without having to go back through the iteration process, send it back through IT, you can open up the viz, make a quick adjustment, and now you have an interactive viz based on a category type. And so I would say that's fundamentally changed the way that we visualize that data.
00:16:31
Speaker
the flip side is i've been in a meeting and you know not too long ago uh... and i and i think i emailed andy and steve the same day where you know that complaints dashboard was pulled up in my company where we could have put it on this flat screen you know touchscreen tv that we have in in our in our big conference room uh... but the person conducting the meeting came in and had
00:16:51
Speaker
literally printed out every page of every drop-down selection so that you know you had interactivity on paper so okay you want to see the fourth drop-down box you know you know flop through four pages and it was just frustrating because you go to the effort of building something that's interactive for those purposes that you could save paper you can interact and be more fluid in the analysis and yet you know people still fundamentally will print
00:17:16
Speaker
Do you think that's because it's sort of what we're used to is working in the print world, you know, maybe not younger people, but those of us are old enough that, you know, we're used to living in the print world and that it takes a sort of culture change aside from obviously the need to have a tablet or a computer in front of you. Because I see this all the time. I have the same experience. So what do you think is the reason why we're not seeing that change or we're still printing out infographics or
00:17:41
Speaker
Like you said, we're printing out all the different drop-down options instead of saying, okay, we're going to come up with a digital solution to this. Yeah, I think it is a culture shift. And our organization as a whole here is a very IT analytics centered culture. So I think it's probably a little easier for us than even other places. But things that we've done is to say,
00:18:03
Speaker
Hey, look what we can do with this interactivity. I mean, put it up on a board and with a touch screen, I can press on this, you know, this, this state and it'll filter by state, or I can select this category and it'll filter by category and being able to be in the flow of that analysis. You know, we use Tableau here at our organization, being able to do that quickly on the fly, uh, then light bulbs sort of go off, right? Cause you can do something that you just can't do on paper. You can move quickly and be in the flow where, where you just can't do that with a piece of paper.
00:18:33
Speaker
jonathan i think in the presentation if you're going to be attending this thing you want the presenter to have curated the findings ahead of time which is and that the m uh... cole uses the uh... metaphor of them shocking clamps horrible oysters to find the pearls and you don't want to show everybody you know all these shells that you shocked you want to show them the pearl and that's great and you know here these things but if in the middle of the meeting someone says
Transitioning from Print to Interactive Dashboards
00:19:01
Speaker
well
00:19:01
Speaker
hate you know i want to see this thing or that thing or this thing that you didn't bring up in this in your carefully hounded and curated presentation having the interactive dashboard jeff alluded to it which is all i don't want to see by this category i want to see by that category because there's something just came up in the news or in our greatest and i'd like to see what's going on there and to be able to do that instantly and dynamically
00:19:27
Speaker
Fantastic. So I'm seeing, I've now been dealing with dashboards for over 10 years. And I'm frustrated that we're still seeing the PDFs are copying into PowerPoint, but I'm seeing a lot more interact activity in the last two years than I was seeing before that. So I'm certainly encouraged. Do you find that people who are, they do have the opportunity, they do have the culture behind them and the organization behind them to have dashboards, do you find that they,
00:19:57
Speaker
offer too many options, too many dropdowns, too many selectors? Or are people more selective about what they allow the user to do? The people who are more experienced are more selective, as opposed to, well, here, you can just filter by every single thing, and then it's just, I'm overwhelmed by this.
00:20:17
Speaker
I had a good story recently where I was working with a large hospital group in the Cincinnati area and one of their reports. They just had this completely filled with drop-down boxes for selection. And they really wanted all of that stuff. They said, we really want the ability to not have to create all of these different reports and we need all of these selection boxes. And so we worked together and said, okay, well, let's have a selection menu.
00:20:42
Speaker
where you can go in on the front end and pick your selections and then see your visualization so that your visualization isn't cluttered with, you know, literally this half a page of drop-down boxes and you don't have any room to show any of the data. So I've seen, you know, I think the future will hold collapsible menus and things like that that we've started to see in Tableau. Somebody's built that out and I'm sure we'll see more of that in the future, but
00:21:05
Speaker
I think the technology will help us in that respect where if we really need to see that level of detail, we don't have to clutter the dashboard with it. Exactly. Or things that are pre-filtered or preset based on a user profile. Oh, this is great. It has exactly what is meaningful to me. But if I want to explore outside of what you think is meaningful to me, I can do so. Yeah, I think in my experience with seeing people who bring dashboards in and they're showing things, I find that
00:21:34
Speaker
I either end up on one end of a spectrum or another, where it's either a PDF report, it's basically a snapshot, and I can't drill down. So it's basically the PDF, it's essentially an infographic, which is fine. Or they go to the totally other end, which is basically the story you just told, Jeff, which is they give me every possible conceivable option, where it seems like a good dashboard lies somewhere in the middle. The creator is guiding you to some point and then allowing you to go a little bit further.
00:22:05
Speaker
Yeah, I would, I mean, in general, I would agree with that. I guess, you know, there certainly are instances where, you know, in this, in this healthcare example that I, that I gave, if they really do need all that data and at least all the drop downs to change the data for so many different things, maybe there's different people that are looking at it. Maybe it's an enterprise level, you know, report that so many people need to see, but they all need to see it for a different reason or for a different area, a different state, a different doctor, whatever it is.
00:22:33
Speaker
I think there's ways that we can give them that through some of the things that Steve mentioned, whether it's user profiles or collapsible menus or auto-filling things based on who the user is. All of those things could help in that situation to wean that down. Jonathan, there's a good example in the book that tackles this. Robert Ruse's multiple KPI dashboard.
00:22:59
Speaker
dealing with a huge organization that had all these different departments that are in all these different needs and they all had their separate KPIs and wanted to devise a way that would allow each department or sub department or sub sub department to focus in on the things that were important without making this overwhelming monstrosity as well as not creating 150 separate dashboards all of which had very similar type of functionality and
00:23:29
Speaker
He came up with a pretty elegant way of handling this, which, where, gee, I don't feel overwhelmed, yet this is meeting a lot of different needs. And his collapsible menu is actually used on another dashboard in the book, the working compensation dashboard, where there's a little button off to the left-hand side, and it collapses the whole left side of the dashboard menu so that all of the drop-downs that are available, all of the options are hidden. Interesting, interesting. Okay, so before we sign off,
00:24:00
Speaker
I want to ask each of you, and I'm guessing there's going to be some overlap here, but what are the top three things people should keep in mind when they're building their own dashboards? Jeff, why don't we start with you? So top three things people should keep in mind.
Tips for Creating Effective Dashboards
00:24:14
Speaker
Well, first I would say the right chart for the data that they're trying to show. I see a lot of misuse where somebody's showing
00:24:22
Speaker
time series data and using something that isn't appropriate for time and it's interesting because there's a whole chapter devoted to the many, many different ways that you could show time. So trying to show something with the incorrect chart type. The next one I would say is use of color. I just find in data visualization in general and especially in dashboard design, people using color incorrectly and so we have a lot devoted to color.
00:24:51
Speaker
whether that be how to encode data when you're using a sequential color scheme to encode a measure versus a categorical color scheme to encode a dimension, something as simple as that, or whether it gets more complex for dealing with colorblind friendly palettes, which we have a number of chapters and discussions in the book on alternative palettes and things that you can do there.
00:25:14
Speaker
And then I guess the last one, number three, would be something just in studying and doing work on my own as I emulate others that I like in their design. The use of white space and just being able to give your visualizations room to breathe, I would say, as we went through most of our visualizations, a lot of it was decluttering exercises, just giving the dashboards a little bit more room, whether that be removing some things that weren't there,
00:25:43
Speaker
that they need to be there or whether it's just redesigning things that are all there in a better way to give the dashboard a little bit more room and don't feel cluttered, those would probably be my top picks. Nice. All right, Steve, your turn. Top three. Yeah, there's a fair amount of overlap, certainly the right chart and color. I've looked through all the dashboards and almost all of them only have two or three colors.
00:26:12
Speaker
And I had just seen some stuff out in the wild and some people making a presentation, and it was just some abominations with misuse and overuse of color. The thing that I might swap out for the white space, although design and clarity and something that just is comfortable to look at, it would come from this, which is get feedback and iterate.
00:26:35
Speaker
The number of iterations these things went through, certainly the dashboards I was building, and I had two great collaborators to give feedback, but also I had to show it to other people as well. And, hey, this isn't clear to this person. It's really clear to me, but it's not clear to my audience. What do I need to do so that they get it? So, iteration and feedback. I would say, I don't know which of Jeff's I would swap out, but I agree with all three of them.
00:27:04
Speaker
But I'd add that fourth. There we go. I think the top four things you need. I think we've got a top four thing. We'll hold on to the fifth one for Andy. Andy can chime in some other time and give us a top five. Andy will say, no screaming cats. No, right? I don't even want to imagine. No, no, you'll see them throughout the book.
00:27:31
Speaker
I've looked at too many of these books on data visualization, and I'm looking at a chart and wondering, does the author think I should make that, or is that something that I should avoid at all costs? And you have to spend two or three minutes to read the thing to realize, oh, that's an avoid at all costs. So we have an icon next to the avoid at all costs, which we call the Screaming Cat. Screaming Cat. So I would trust then when the audio book comes out, we'll have a Screaming Cat audio.
00:27:56
Speaker
screech that will show up in there. Yes, we will be auditioning people to do the Screaming Cat. So there you go, folks. If you're looking for some freelance work, you can call Jeff and Steve and Andy to audition for the Screaming Cat for the audio version of the Big Book of Dashboards. Steve Lexler, Jeff Schaper, thanks so much for coming on the show. This has been really interesting. Thanks a lot. Thank you, John. Thank you so much, Jonathan. It's been a pleasure.
00:28:21
Speaker
And thanks to everyone else for tuning in to this week's episode. If you have questions or comments about the show, please let me know. And of course, do check out the Big Book of Dashboards. The link is on the show notes page. So until next time, this has been the policy of this podcast. Thanks so much for listening.