DataMastersPodcast

DataMastersPodcast

Episode 10 — released August 26, 2020 • Runtime: 28m17s

The role empathy plays in data visualization

Nick Stepro

Nick Stepro

Nick Stepro, Senior Vice President of Product Management, Arcadia Healthcare

A key part of data is communicating it to others, and that’s where data visualization comes in. Nick Stepro, senior vice president of product management at Arcadia Healthcare, has made data visualization his career. He talks about how he got into data visualization, why you need empathy in design and how he manages to communicate the value of data without taking away from its meaning.

Transcript

Nick Stepro:
It’s not just to communicate data efficiently, but it’s to grab an audience that otherwise would not look at this and to force them to engage in that material.

Nate Nelson:
Hey everybody, and welcome to the DataMasters podcast. My name is Nate Nelson. I’m sitting with Mark Marinelli from Tamr. He’s going to introduce the subject and the guest of today’s show. Mark, how’s it going?

Mark Marinelli:
Pretty well. How about you Nate?

Nate Nelson:
Good.

Mark Marinelli:
All right. Let’s jump in. A key part of data is communicating it to others, and that’s where data visualization comes in. Nick Stepro, Senior Vice-President of Product Management at Arcadia Healthcare has made digital… Ah.

Mark Marinelli:
All right. Let’s go. A key part of data is communicating it to others, and that’s where data visualization come in. Nick Stepro, Senior VP of Product Management at Arcadia Healthcare has made data visualization his career. Today you’ll hear him talk about how he got into data visualization, why you need empathy in design, and how he manages to communicate the value of data without taking away from its meaning.

Nate Nelson:
Here’s my interview with Nick.

Nate Nelson:
In data we have these big unanswered matters of how to gather the huge amounts of information in the world, how to make sense of it, and how to make use of it. With such big questions the subject of data visualization that we’re talking about today could potentially come across as a bit small. Why isn’t that the case? Why is it worth dedicating a whole podcast episode like this to the subject of data viz?

Nick Stepro:
That’s a great question. I think the best way to think about this is to reframe the narrative of data visualization away from one of sort of technical minutia of color-coding your graphs to one of just basic communication. That’s all it is, plain and simple.

Nick Stepro:
If you think about how we intend to use data culturally within the workplace it’s to take a object of information from one place and to jam it into a human being’s brain. If you think about those brains, they’re just these big, beautiful processors that are capable of a ton of things, of pattern recognition and drawing from experience and context.

Nick Stepro:
So the question is if you have all of this data you’ve been curating and cultivating and wrangling how do you get that into the processor that can do something with it, and that’s an exercise in communication. That’s what data visualization is, and I think without that the data is largely useless.

Nick Stepro:
I think of this sort of story, and it happens to me a lot. It may happen to you, this concept of like the finance meeting, right, the exec team gets together, the meeting is more expense than God. You’ve got 13 people flying in, spending a few hours together, and the finance team comes to the meeting and presents really critical information in some massive spreadsheet.

Nick Stepro:
It’s got like 72 columns. You’re already on the double letters in Excel, and you’ve got 15 people around a table trying to figure out what the heck it means, and then 10 minutes in they point you to cell AL72, and it’s like that’s what matters. That’s the field that means something.

Nick Stepro:
This happens a lot. What a more efficient way it would be if we communicate that information visually in a way that these beautiful brains around the table could quickly contextualize that data and make decisions rapidly rather than spending our time trying to parse through piles and piles of stats, right?

Nick Stepro:
That’s the big picture on data viz. It is less an exercise in minute technical details and more one of just basic communication of the information you have.

Nate Nelson:
Great. I want to get into some of the details of what you just said, but firstly how did you land on this specialty, working not just with data, but visualizations in particular?

Nick Stepro:
I got here largely by accident. I think I started in basic web development and studied econometrics and stats and things like that. I happened to enter the workforce at a time where web development was beginning to support data visualization in a wholly unique fashion and enabled us to sort of pull our heads out of Microsoft Excel, bar charts and pie charts and all that stuff, and think more freely about how you visualize information, particularly how you visualize in an interactive format.

Nick Stepro:
That confluence of those two things and the fact that I operate in healthcare, where there are really unique challenges in information design, information presentation, brought me to this place and I haven’t looked back for largely the reasons I’ve just mentioned, the level of importance of communicating data across a broad spectrum of consumers.

Nate Nelson:
So the first most obvious purpose of data visualization that comes to mind is to translate complex ideas and numbers into a simpler form for dummies like me. Is there more to it? What’s the fundamental purpose of doing visualizations?

Nick Stepro:
I think there are many, and I’ll get into that in a second, because the reason you visualize data can vary. It can be towards an artistic pursuit. I can be towards an efficiency or a power user angle. It can be for data journalism, so for broad communication of simple concepts to folks.

Nick Stepro:
But the analogy I most like to think about is that of the hard drive to the processor. We’ve shoved all of our data up into the cloud, and that’s all well and good, but it’s only as good as it can be communicated down to local processing units, right?

Nick Stepro:
I think the same is true on the data front. We have spent untold money and effort consolidating massive swaths of data, but that data is only as good as the bandwidth it and human processors. So you have this challenging computer science that you’re probably aware of, this problem being IO constrained, right? You’re not constrained by the amount of computational power that you have. You’re constrained by the ability to read and write data to and from those computational units.

Nick Stepro:
We run up into that problem a lot in the data world, where you’ve got people capable of making a powerful decision when armed with the right information, but it’s very, very hard to efficiently get that information to and from their brains. So in order to get over that IO constraint we focus on expanding bandwidth, and that’s exactly what data visualization is.

Nick Stepro:
It’s broadening the pipe so that in a shorter period of time a human being can contextualize information and make decisions on it, and there are a bunch of ways to do that. Visualization is only one of them. It just so happens that human beings have been wired in this way that responds to visual aspects better than a lot of other things, and so presenting quantitative information in a two-dimensional plan or a three-dimensional space so that we can interact with them sort of as objects. That goes a long way in accelerating our understanding.

Nate Nelson:
One of the terms I’ve come across in speaking with you recently is this notion of empathy in design. Nick, what does it mean to have empathy in design?

Nick Stepro:
Yeah, it’s hard to overstate the importance of this. It’s really the only thing that matters, right? Again, if you take this lens that data visualization is an exercise in communication, how do you effectively communicate to someone if you don’t truly understand them or you don’t attempt to deeply understand them?

Nick Stepro:
My day job is not just data, but is running a product development team, right? So in the product world we have this concept of human centered design, where instead of just delivering features you attempt to profoundly understand the needs and problems and barriers for the users that you’re designing for.

Nick Stepro:
The same is true in data visualization. It’s not sort of a one size fits all exercise. You need to think about who you’re communicating with and what their background is both culturally, education, their level of familiarity with the subject matter.

Nick Stepro:
There’s this problem that you may understand. It happens to me a lot. It’s the challenge of like expert bias, where someone creating a data visualization has been sort of knee deep in some piece of data or domain for hours, days, weeks, months, and they over estimate the level of expertise of their target audience and they sort of make leaps in their data design so that fresh eyes when they see it are not able to contextualize in the same way that the author is.

Nick Stepro:
There are ways to combat that. One of them is to have friends and shop your designs around early. When you do that, you have to avoid… Earlier on in my career I would always have this gut reaction, you show sort of a piece of work to someone and they’re like, “Well, I don’t understand. What’s the blue color mean?” It’s like no, you idiot, the blue color means this. You sort of have this harsh reaction.

Nick Stepro:
You need to avoid that. You need to welcome the fact that people are coming at it from a different set of backgrounds and levels of understanding and take that feedback and then refine your design so that as you shop it around to further individuals they’re able to get the value out of it that they want.

Nick Stepro:
There’s a couple of other useful tips. The concept of the legend, I don’t know if you… If you look at a lot of data visualizations they can skew from very utilitarian to very artistic. I happen to lean towards the aesthetic end, so the concept of adding a whole bunch of scales and legends is kind of frustrating, because it can impede the aesthetic pursuit of that piece of work.

Nick Stepro:
I’ve only recently leaned into them and sort of accepted that legends can be beautiful and that there are really beautiful artistic ways to draw a consumer in and to help them understand the information that you’re trying to present while still maintaining some of those more endearing aesthetic pursuits that we have.

Nate Nelson:
One of the other ideas that’s bounced around is this notion of accessibility versus power users. I want to bring your attention, Nick, to one of the many visualizations on your personal website. This one in particular is titled I Love Nick Cage. Now for listeners who can’t see what I’m referencing, it should suffice to say that this visualization in particular has so many notes and so many connectors that it nearly destroyed my computer.

Nate Nelson:
So, Nick, explain yourself. Why torture me by taking something I thought I knew and making it look like rocket science? I thought this was the opposite of what data visualization was?

Nick Stepro:
Yeah, yeah, yeah. Part of that is… My personal site, is a reasonably tongue in cheek attempt at visualizing mundane information in overly complex ways, so part of its artistic pursuit is to misdirect to some degree, but I think that’s a great jumping off point.

Nick Stepro:
What you’re referencing is sort of a force directed tree style diagram, which is topical for this podcast. If you go to basically any big data shop, any startup that exists or existing big data shop, they usually have some background image on their header that has a bunch of nodes being connected by all of these lines, and that’s like the best we have at visualizing big data, because it makes people think that there are these fancy neural networks making all of these inferences, et cetera, but it’s just a simple force directed diagram, which is a way to present co-occurrence relationship information, et cetera.

Nick Stepro:
In the case of Cage, who happens to be one of my favorite actors, he has spanned a career of where he interacts with hundreds and thousands of other artists, so in a relationship diagram like that we can show an individual and span out sort of in a six degrees Kevin Bacon kind of thing… Fan out and understand how those relations have built and how multiple movies have been shared between actors exist.

Nick Stepro:
Again, it’s kind of tongue in cheek. I think the concept of design considerations across different types of audiences is actually a really important topic. Again, if you anchor data visualization as just a means of communication you can think about an alternative means of communication, which is the written word, and when you think about the written word there are many different modes by which that’s delivered.

Nick Stepro:
One of them would be poetry, where the intent of that written word is kind of ambiguous, right? It’s attempting to be meditative. It’s trying to appeal to potentially emotion rather than just pure reason or deeper philosophies.

Nick Stepro:
Then you have textbooks, also written word. They don’t make a lick of sense to anyone that reads them except for those that have a certain baseline of understanding, but they’re well indexed, they’re data dense. You can quickly download a tremendous amount of information.

Nick Stepro:
Then you have things like public signs, street signs, right? That’s also written word, but they’re explicitly… At a simple reading level they’ve got these big fat round Helvetic letters, iconography, et cetera.

Nick Stepro:
Those are all different modalities of the written word, and if you think about data visualization there are analogies for all three of those, right? For the textbook you have these… You know, if you’re familiar with these dashboards in Tableau or Power BI that have a thousand filters and slicers and dicers drill downs, it might not make any sense to somebody outside of that domain, but for someone that knows what they’re doing it’s really a quick method to interact with information and gather insight.

Nick Stepro:
On the poetry side you’ll see on my site and on the data gallery that I maintain at Arcadia a lot of those pieces skew towards art. They’re not explicitly designed to quickly communicate information. There’s a piece on there that I like I made a few years ago. It’s called the Final Year. We’re looking how we die in America. As an aside, the answer to how we die in America is that we do it really poorly. We’re really bad at dying. We don’t plan for it. We throw money at the problem. We die in hospitals instead of at home.

Nick Stepro:
It’s a topic that I feel really strongly about, so instead of just communication that stat, which frankly could be communicated in a simple bar chart, I wanted people to meditate on the topic, so we created a very abstract piece. It actually looks like a waterfall of millions of blips of data, one piece of data representing a visit between a patient and a provider in their final year.

Nick Stepro:
It’s technically like a scatter graph, where every data point is actual XY access, but it is presented in a way that is not immediately obvious and you really need to sit and meditate on that, and there’s value in that, the same way there’s value in poetry.

Nick Stepro:
Then there’s the public sign metaphor. The intent is to communicate to folks that don’t have time or have a broad baseline of understanding, and it’s a totally different kind of information display. Think about like your classic executive dashboard. Someone has five seconds, they want a 30,000 foot view. It needs to be boiled down, all of that jargon, right? That’s a difficult exercise in and of itself. We’re not appealing to someone that’s going to invest a tremendous amount of time in something.

Nick Stepro:
I’m working on a piece right now and it’s going to go out to hundreds of healthcare executives, so you do what you’re supposed to do. In this type of topic you focus on primary colors. You’ve got stop light metaphors. You don’t throw in a ton of spark lines. You keep it really, really simple.

Nick Stepro:
You don’t want to alienate the user. A lot of these healthcare executives, if they see something they don’t understand they feel alienated, they throw the report book out, right? So you want to be very friendly and approachable.

Nick Stepro:
But that comes with this really, really big trade off because you reduce the resolution of the data and you open things up to misinterpretation. One of the challenges that I’ve come up to is boiling things down to a letter grade. How do you really communicate simple stuff visually to somebody… Quantitative information, but communicate it simply? One of the ways to do that is just to default to a letter grade, right? I’m not going to tell you all of the details, but someone got an A, B, C, D or F.

Nick Stepro:
So I built that out, but then you stumble upon another challenge, and this gets back to the empathy in design, which is like are the grades curved? Is the C average? Most of these healthcare executives probably never got lower than a B minus in their life, right? So all of a sudden you show a C on there and people get fired because they think it’s a F since all of a sudden you’ve tried to boil things down and create a visual paradigm that is approachable and simple, but you put a tremendous amount at risk in doing so, if that makes sense.

Nate Nelson:
This brings me to the question of creative input. You, Nick, get to choose what makes it into a visualization and what form that visualization takes. I rarely have the thought, going through your website, that oh yeah, this is exactly how I would have mapped out this information. It’s usually more like how in the world did he come up with this? What goes into making these decisions about what information to include and also how to represent it?

Nick Stepro:
Honesty it’s an exercise in breaking your prior understanding of visualization. The concept of a bar chart and a line chart, they’re relatively new concepts, right? So the way people used to describe quantitative information was a little bit more freeing, if that makes any sense, because people weren’t confined to using Excel and PowerPoint.

Nick Stepro:
It wasn’t just like a button that you hit to generate a pie chart. You had quantitative information and you had a lot more freedom in how to display. We now have a little bit more of that freedom, so part of the creative decision making is a little more organic when you free your brain from the confinement of the bar chart and the pie chart. When you think about data being able to represent any array of shapes, lines or objects, then you can really let the data do the talking.

Nick Stepro:
I often think of my own brain as a computer processor and I look at a set of data and I say what would be interesting ways to try to expand my understand my understanding of this data and feel out the different shapes and textures of a given piece of data? Often that leads to some pretty wild outputs.

Nick Stepro:
To your point, oftentimes the most efficient way to communicate something is with a series of bar charts, but that is not the sole pursuit of a lot of that data gallery and a lot of data visualization in general. It is not just to communicate data efficiently, but it’s to grab an audience that otherwise would not look at this, I think to force them to engage in that material.

Nick Stepro:
There are good ways to do that, and a lot of them are to show colors and shapes and outlines that someone has never seen before, because they’re immediately drawn into what that might represent.

Nate Nelson:
Now that we’ve talked about the theory and the practice of it, how can data viz affect business outcomes?

Nick Stepro:
I think in many ways it has been sort of like cheaply applied. If you think about 10 years ago, just like the explosion of the infographic. I don’t know if you remember this, where every company was putting out these kind of bogus, cutesy infographics with pseudo quantitative information. That’s the angle where I think folks have leaned on this medium from a pure marketing standpoint.

Nick Stepro:
I think it can go off further. I think we are… Data literacy amongst consumers and businesses has been expanding considerably, and I think the extents to which a business can embrace that and meet their users at a higher level of literacy and provide meaningful ways to interact with data… I mean you think about any modern technology company is really a data company. There are very few that aren’t in this in the pursuit of building a set of data that we can then create inferences from and build platforms off of.

Nick Stepro:
With that in mind, if you’re sitting on this wealth of data and you’re not thinking about how to intelligently render it to both your internal stakeholders within the organization but also externally to your customers, you are leaving a ton of value on the table for that information.

Nate Nelson:
Is data visualization a science or an art?

Nick Stepro:
It’s both. Some people sit on different ends of that divide. There is real science and behavior studies and cognitive studies that go into how to present quantitative information to the human brain. Again, thinking about that broadband channel between hard drive and processor, there’s very real biology behind how the cortex interprets a bunch of visual stimuli and converts that into knowledge, so there’s a tremendous amount of science that goes into that.

Nick Stepro:
I skew, and you’ll see this in a lot of what I do, probably more towards the art standpoint. I think we have these tools now to be very expressive with data that we have. Our whole world is data, right? So when you watch a movie that’s data. That’s columns and rows where every pixel is a color that represents something and the mosaic of them together represents a world that we see.

Nick Stepro:
It’s not dissimilar from how you think about rendering data. You can create these beautifully rich pictures that appeal to both the emotion and the reason of individuals. That’s what gets me off in this space.

Nate Nelson:
Nick, to finish things off here, why do you feel passionate about data viz?

Nick Stepro:
I think there’s a selfish reason and then there’s a bigger reason. On the selfish reason, I think there are very few disciplines that let you combine so many parts of your brain and experience into one pursuit. I love crunching numbers. I may spend my weekends pounding Python and JavaScript and Sequel and all of that. I’m not very good at coding, like the code is hideous once you look at it, but I love interacting with technology and data.

Nick Stepro:
I also really like communicating and I like thinking about human beings and how we want to interact with the world around us. I like aesthetics. I spend a lot of my time browsing design blogs and thinking about the evolution of design outside of the world of data visualization.

Nick Stepro:
I think it’s just very cool that I now live in this time where all of those concepts can converge into a single pursuit. It is like candy to your brain when you’re able to sit down and solve a problem that draws from that many areas. So that’s the selfish reason.

Nick Stepro:
The less selfish reason is largely what I was describing before, that we live in a world where we are now capturing… The amount of data exhaust that we generate and is being captured is immense and growing at a crazy rate, so we have this new quantifiable lens of what humanity is. The best way, or one of the best ways that we know of how to communicate that complex information to each other is visually, via data viz.

Nick Stepro:
So I think there’s a tremendous future in the pursuit of harnessing this wealth of data that describes or approximates humanity in very different compelling ways, figuring out how to render that to further our understanding of ourselves and each other.

Nate Nelson:
That was Nick Stepro. I’m back with Mark Marinelli. Mark, any thoughts on what you just heard?

Mark Marinelli:
So many of our other guests have talked about the work required to prepare data so that companies can draw meaningful results from it. Nick offers a great take on what happens after that work is complete and you have your datasets, that sort of last mile of visualization. You need to present that data so that people can understand and use it. His story serves a good bookend of what’s been discussed in the earlier episodes.

Mark Marinelli:
Putting what Nick said into the greater context, telling a story with data is a detailed process. From getting the right data, to cleaning it up, to visualizing, a lot of work goes into creating a chart or a graph, and providing your users all of the context necessary for them to be successful is extremely important.

Nate Nelson:
All right. With that, thanks to Nick Stepro for speaking with me, and thank you Mark.

Mark Marinelli:
Thank you Nick. Take care.

Nate Nelson:
This has been the DataMasters podcast from Tamr. Thanks to everybody listening.