Kirill: 00:00:00
This is episode number 395 with the author of Storytelling With Data, Cole Nussbaumer Knaflic.
Kirill: 00:00:12
Welcome to the SuperDataScience podcast. My name is Kirill Eremenko, Data Science Coach and Lifestyle Entrepreneur, and each week we bring you inspiring people and ideas to help you build your successful career in data science. Thanks for being here today and now let’s make the complex simple.
Kirill: 00:00:37
Welcome back to the SuperDataScience podcast everybody, super excited to have you back here on the show. In today’s episode, we’ve got a very exciting guest, none other, but the author of the famous book, Storytelling With Data, Cole Nussbaumer Knaflic. So in case you haven’t heard of Cole before, Cole is the author of one of the most popular books on data visualization, is called Storytelling With Data. It is a number one bestseller on Amazon, is got over 700 ratings and it is the most recommended book on this podcast. So finally, we had Cole on the show and I literally just got off the Zoom call with her. In today’s podcast you will hear about quite a lot of things. First of all, you’ll hear about Cole’s story and how she worked at Google and how she started her own business, Storytelling With Data and what she does at present.
Kirill: 00:01:36
Then we talked about the top three tips for data visualizations. These are tips that you can really take away today and start applying in your visualizations. Also, you will learn why data visualization is actually so important. In fact, it’s actually more important than the data modeling side of things and why you should be an expert at it if you want to be successful in data science. Also, we’ll talk about why it’s important to have an opinion on the visualization that you’re creating, how to voice that opinion in your visualizations. We’ll talk about storytelling with data. In fact, this is the first podcast on this show where we dive deeply into storytelling, and we talk about things like a narrative arc and how that approach is different to what we see typically in business, which is a linear style presentation, and how to shift from a linear presentation to a narrative arc and what are the steps involved with that.
Kirill: 00:02:26
You’ll hear about the number of steps in a story. Cole will share an amazing strategy for building up complex visualizations, if they have to be complex, and then we’ll have a lightening round of questions with the questions that our audience submitted on LinkedIn. So a very cool podcast coming up, data visualization is an extremely important part of data science and whom else better to learn from than Cole Nussbaumer Knaflic, probably nobody. So without further ado, I bring to you the author of Storytelling With Data, Cole Nussbaumer Knaflic.
Kirill: 00:03:07
Welcome everybody back to the SuperDataScience podcast, super excited to have you back here on the show. Today, we’ve got a very special guest, Cole Nussbaumer Knaflic calling in from Wisconsin. Cole, how are you?
Cole: 00:03:17
I’m great. Kirill, how are you doing?
Kirill: 00:03:19
Awesome. So excited to have you on the show, big time fan. Got both your books. If anyone’s watching the video right here, both of them, this is so cool. Tell us, so Wisconsin, how long have you been in Wisconsin for?
Cole: 00:03:32
Yeah, it’s gone by fast, but we’ve been here just about two years now, Bay Area Transplant. My husband grew up in this area and his side of the family’s here, so we were ready to get out of San Francisco and we’ve got three little kids and wanted a little more space and to get closer to family.
Kirill: 00:03:49
Nice.
Cole: 00:03:49
And so that’s been a good transition.
Kirill: 00:03:51
Nice, fantastic. And I was watching one of your talks, so back at … is it in California where you worked at Google?
Cole: 00:04:01
Yes, in California.
Kirill: 00:04:03
But since then, now you have your own website, your own business. So I guess it’s all virtual, is that how it works?
Cole: 00:04:10
Today, it is all virtual. Historically, that was not the case. So my company is called Storytelling With Data and we spend our time helping people and organizations make graphs that make sense. And so this primarily takes the form of workshops where we’d go into an organization, spend half a day or a day with a team really covering foundational principles for both showing your data, but then also communicating to others with that data and thinking about how you weave data into a narrative, into a story that helps drive action. So historically we would go to organizations mainly and do that, though we’ve shifted to the virtual world, that many of us are spending our time in these days. And so we’ve brought what has historically been in person interactions into the online environment. So again, doing workshops for organizations and then also for individuals who want to hone their skills.
Kirill: 00:05:07
Okay. And how’s that been going for you for transitioning into virtual?
Cole: 00:05:12
It’s actually been going really well and I historically was anti-virtual. I have this feeling that there’s so much gained from the in person interactions and from being able to be there and physically move around a room and see people’s faces to know whether folks are getting it, or if we should pause and go more into detail. And so I’ve been delighted by how well virtual environment works for this stuff. And what we’ve done is we’ve tried to be really creative in how we’re structuring that in terms of how do you let people interact? How do you get people working on things directly? And we’ve learned a lot from the process and some surprising things.
Cole: 00:06:01
So for example, we get a lot more interaction from people who would consider themselves introverts, who maybe wouldn’t have spoken up in a group setting in the same way, and virtual scales a lot better. So if I’m presenting and I’m talking at the camera to the audience, every person feels like you’re talking to them. So it actually, and this surprised me, but it can feel like a more intimate learning environment than being physically there because you simply can’t do that in a room with 30 people. And so there’s been some good surprises and yeah, it’s been fun to … because we haven’t been in this mode of creating things new and starting from scratch on some things. And so it’s given me and the team opportunities to try things out and test things out, iterate very quickly and learn from what we’re doing in ways that I think has been energizing.
Kirill: 00:06:52
That’s so cool. Absolutely agree with the notion that, presenting virtually. I was also presenting a few months back and I had to think about this, that the person watching the screen is like one person or maybe like two, if it’s like a colleague or a friend or a family member, but that’s it. It’s like, it’s very different. You don’t have that audience effect, so even the words you use, it’s not like welcome everybody, it’s like welcome you, individually.
Cole: 00:07:23
Yes.
Kirill: 00:07:23
So yeah, that’s so cool. How did you get into this space? How did you get into data storytelling? It’s like such an interesting area to be in.
Cole: 00:07:32
Yeah, it’s super interesting. So I studied my undergrad is in applied mathematics, so I always enjoyed the puzzle piece of math and the logic and making things fit together in ways that made sense. And then out of university, my first job was in banking, in credit risk management, building statistical forecast models to predict when people weren’t going to pay or how much they’d default, so that we could set our reserve levels. And I think that for me is where the seeds of this all started, because I would spend a lot of time designing my graphs. And so I did quite a lot of presenting to senior management and I would spend time on the details of the graphs and things like color.
Cole: 00:08:16
And at that point I was just trying to make things look nice, right? Without knowing really anything about the underlying principles, but the reinforcement that I got early on through doing that was I recognized that people were paying more attention to my work as a result, which reinforced this wanting to spend time on the visual, on the aesthetics, on the communication piece, which was something that my colleagues didn’t typically spend as much time on. Went to business school while I was in the banking world and then the subprime crisis happened. Banking was not a fantastic place to be in and so I thought, well-
Kirill: 00:08:52
You didn’t cause the subprime crisis, did you?
Cole: 00:08:55
I did not cause it, no. Actually in credit risk, we were the ones that were saying, whoa, whoa, whoa, we’ve not been in this territory before. We can model things, but you can only model when you’ve got the right sort of history to model off of. And we were doing new things with no documentation and all this crazy stuff. But the salespeople and the volume of money coming in made that a hard fight to fight until everything crashed. Right? And then you realize like, that’s what they were saying ahead of time. But anyway, I wanted to get out of banking and so did some self assessment at that point to say, what skills do I have that I want to be using on a daily basis? What sort of company do I want to be at, where do I want to be? What skills do I want to develop? And from there ended up at Google on the people analytics team, which is, it’s analytics team-
Kirill: 00:09:42
Hold on, sorry, I got to interrupt here. That is so easy to say, I ended up at Google. Google has a seven stage interview process, how did you end up at Google?
Cole: 00:09:52
I submitted my resume, my application online. So I’d been looking at different jobs and I was looking everything from quantitative marketing sort of roles. And then I just, I came across this description for people analytics, I thought, that sounds really strange, analytics on people? That could actually be sort of interesting. And it turns out the same sort of stuff I was doing in banking applies in this totally different sphere. Right? Cause if you think about it, modeling attrition, someone to leave the organization is actually very similar behind the scenes to modeling when someone’s not going to pay your loan. And so it got my brain going in that direction. And then yeah, I mean, I was very lucky in terms of millions of, I forget what the numbers are, but it’s crazy the number of applicants that Google got at that point.
Cole: 00:10:41
And I’m sure that number has increased now, that somebody picked my resume out and said, let’s talk to her. And so I had the [inaudible 00:10:48] phone screen and the in person’s and all that jazz, but actually went pretty quickly through the process. And yeah, a couple months after applying was moving to California and getting used to like biking around Mountain View and had this dream job, right? Because it was when I started was right as the people analytics team was formed. So prior to that, there’d been a couple analysts and compensation, a couple analysts and benefits. And so they had hired Facade Seti, who was my boss to bring that together, run the analytics organization within people operations or human resources. And yeah, it was that were looking, every data we looked at at the beginning was data that nobody had really looked at before, which was a really exciting time to be learning about people through the data and what we could do with that information.
Kirill: 00:11:38
Are you able to share some insights about what objectives you had on a day to day basis? How you were helping Google with their people data?
Cole: 00:11:48
Yeah, it really ran the gamut, right? So like any organization, when you’re first starting to get your around data, it starts by figuring out well, what do we have, what can we learn from what we already have at our fingertips? Where do we collect more data? So at Google, we would do massive employee survey every year. Questions about how you feel about your manager or the work environment and different things that we would be able to use to change directions or build strategy around how people were, what’s the right word? Like I think Google was interesting because there is such value placed on people and there was good recognition from the founders and all the way down that it’s the people, it’s the brains that make all of the great products.
Cole: 00:12:37
And so by doing things to keep them happy and wanting to be at work and feeling supported by their managers and all these little things that add together to make up this fantastic work environment where it doesn’t feel like you’re working because you’re working on cool stuff and it’s sort of fabricated so that you never want to leave, which shifted over time as the employee population aged and started getting married and having kids, all of that sort of stuff. So I know things have changed since then, but it was a neat time to be there for sure. And also had the opportunity while I was there because I had still paid a lot of attention to how I was showing data.
Cole: 00:13:16
And at one point we were building out an internal training program and people building it came to me and said, “Hey, data visualization, you like this space, can you build us some courses focusing on this?” And so that I thought yeah, that’d be fantastic. And so it gave me an opportunity to pause and do some research. This was like 2009-ish, find Steven Few’s books and Edward Tufte’s books, which were like the main things out there at that point. One of the things that I think has changed dramatically and when the decade plus since then is just access to information. There’s so much more, so much more readily available now in this space than there used to be. But so developed some courses on how to show data effectively, initially for this training program. And then we ended up having broader interests so we rolled it out across all of Google.
Cole: 00:14:08
I got to travel around offices in different parts of the country, different parts of the world and teach people what I’ve learned, which was super exciting, and then started getting inquiries from outside of the companies. And hey, we’ve heard you do this, can you come do this for us too? And so eventually decided it was time to take this show on the road, if you will. And recognizing that these are not just skills needed by Google, right? These are skills that anyone can use to have greater impact when it comes to how they communicate their technical work.
Kirill: 00:14:44
And that’s very admirable. I think, from your LinkedIn, I saw that you stayed at Google for quite a while. It was like five, or five or seven years, something like that.
Cole: 00:14:55
I was there, let’s see, from 2007 to 2013. So yeah, six years.
Kirill: 00:15:00
Six years, then clearly it shows that you are enjoying what you’re doing. But then like to take that jump and jump into that uncertainty and go like, oh, okay, the world deserves to know about these techniques and start this business and help. Starting a business it seems to an outsiders that it’s easy. It’s much easier to keep your amazing job at Google than to do this. So hats off to you for that. And thank you because your work has impacted me. And I know lots, your book, as I mentioned before the podcast, is the most quoted book on this podcast. So speaking of the book, at what point in time did you decide to write the book? Oh, by the way, for those who don’t know, it’s called Storytelling With Data and there’s a second one Storytelling With Data: Let’s Practice. Totally, totally love it. We’ll talk about it in a second. When did you decide to write it?
Cole: 00:15:50
So I had been teaching workshops, so left Google, had been teaching workshops at different organizations and learning a ton, right? Because before going into any organization, I would collect examples ahead of time to get an understanding of how the group was communicating with data currently, we’d use them as the basis for some of the hands on exercises that we would go through. And it was always very interesting to note that it’s sort of irrespective of industry or the part of the organization, everyone struggles with the same sort of things. And there are common, relatively easy things that we can do differently in ways that just enable us to get our information across quicker and with fewer misunderstandings and get people to really focus on the so what, right? Instead of like going down the death spiral of wanting more data, more data, more data.
Cole: 00:16:43
And so that was an interesting recognition as I started doing more work with more companies across more industries, is just the pervasiveness with which people could apply these lessons. And so it was after being a couple years into doing it on my own that I realized I wanted to codify the lessons in a book. My husband will say, well, I had the idea for a book way longer ago than that, but I wasn’t ready to write it yet. It took going through and you have to be ready to write a book because a book is a huge commitment and a huge time investment.
Kirill: 00:17:19
I can imagine.
Cole: 00:17:19
Because it took a little over a year from, by this time I decided, okay, I’m going to do this to finishing writing. It was probably the better part of a year, a little bit more. But so that was probably, let’s see, it came out in November 2015. So that was 2013 or so sometime when I started writing the book and pulling together lessons and identifying examples. And the main goal with that was just to help more people learn, right? And to point out the things that I had learned, because me standing in front of a group, giving a workshop, you can only talk to so many people, but a book can be out there and find its way to all sorts of corners of the world, which has been super amazing to watch.
Cole: 00:18:02
And then Let’s Practice came quite a bit later. And it was funny because when I got done with the first book, I thought, I will never have another word to write again, right? I’m out of words, I’ve used them all.
Kirill: 00:18:11
I know the feeling, yeah.
Cole: 00:18:14
But that goes away, right? Over time, you’ve been talking through something, you’ll find a new way to explain something. Like oh, that’s good. Right? I should use that somewhere. And so by the time I started writing Let’s Practice, which I wanted to be much more of an experience. And one of my fears is that people would read the first book or go to a workshop and they get excited and they get it. But that, it’s very easy to just go back to doing things the way that you’ve always been doing them. And so wanting to give folks a reason to practice and some low risk ways to do so where you can try something out, right? The data’s provided, the scenarios provided, there’s no risk to you, but you can test out different things and learn from that process. And so that’s really where Let’s Practice came in. And when I initially started it, I thought it was going to have … I thought there’d be more writing in it, that I’d have a chapter and then some exercises.
Cole: 00:19:04
But I started writing and realized I had way too much to say for that, it was going to become this like thousand page book and so scaled back and thought, well, I’ll do the lessons that I want to cover through the exercises. And so Let’s Practice is organized into three sections within each of the chapters and the chapters follow the same general order as the first book. So it’s like read the first book or start with a chapter of interest and then turn to Let’s Practice. In the Let’s Practice you’ll find three different sets of exercises. There’s practice with Cole, where I pose a scenario and some data that you’re meant to solve on your own. But then I also show you how I solve it and have the thought process behind it. And this is where the bulk of the content of the book comes in, so you’re really learning through example.
Cole: 00:19:49
And by getting this window into someone else’s thought process and constraints and considerations, when it comes to these sort of off the wall examples. Because I felt like in looking back at the first book, everything was very simple. Everything was very cut and dry, which is a great place to start, but the real world is messier than that. And so with lots of practice had the opportunity to introduce more of that messiness, and address ways of thinking about that. And then the second exercise section is practice on your own. So its similar sort of canned examples, but without any predefined solution. So this is great for the individual who just wants more practice or university instructors or other instructors who are teaching the material where you can draw on for tests or assignments or group projects or for the manager who just wants their team to practice a little bit more.
Cole: 00:20:39
And then the final exercise section within each chapter is practice on your own. That’s okay, you’ve done this in theory, right? You’ve done it with some canned examples. Now take a project that you are facing in your day to day, and here’s how you break it down. Here are the questions to ask, here’s where to get feedback and who to get feedback from and all these really practical exercises that you can do with your own project in mind, that will step you through different aspects of communicating in ways that will get your message across, get it heard, get it understood by your audience.
Kirill: 00:21:12
Love it. This episode is brought to you by SuperDataScience, our online membership platform for learning data science at any level. We’ve got over two and a half thousand video tutorials, over 200 hours of content and 30 plus courses with new courses being added on average once per month. So all of that and more you get as part of your membership at SuperDataScience. So don’t hold off, sign up today at www.www.superdatascience.com, secure your membership and take your data science skills to the next level.
Kirill: 00:21:51
Love it, and the books, what I love about the first one, Storytelling With Data is like, it’s so easy to read. It’s got a lot of pictures, it’s got very … Like you’ve mentioned it takes a year to write a book, but that one of the challenges is to find a way to break it into such a manuscript that is simple for people to understand.
Cole: 00:22:13
Yes.
Kirill: 00:22:13
And second one, great compliment in terms of actually doing the practice. Now that we’re on the topic of the book, you probably get this question, asked this question a lot. What are your, let’s dive into the content, right? So lets talk about some best practices of visualization. There’s a lot in the book, from pre-attentive attributes to short term memory, to cognitive load and stuff like that. Very interesting also from a psychological point of view, like what is the psychology of humans behind this? If you had to take the top three tips from the book and share it with somebody to already start using today, what would they be?
Cole: 00:22:54
Yeah. So three things I would say when you are communicating with data, so first off is think about how you can simplify and declutter. And when I say simplify, it’s not about oversimplifying rather, it’s not making things more complicated than they have to be, because it’s actually, it’s quite easy to take something simple and make it appear very convoluted and complex, that does bad things when it comes to the efficacy of the communication. It’s a much harder thing to take something that is truly complex and communicate that in a way that’s easy to understand, and that’s what we want to aim for. And oftentimes we, because we know our data so well, right? We’ve spent a ton of time with it. We know what we want to say with it. We know where we think people should focus. We know what’s there that may not be relevant or necessary that you can kind of ignore, but the challenge is for anybody else, they’re less close to that data.
Cole: 00:23:50
They maybe are looking at it for the first time. And so we have to take explicit steps to make those things clear to others. And so simplifying, decluttering, getting rid of anything that doesn’t need to be there or that isn’t impacting the message, right? You might have done a hugely technical statistical analysis, but depending on your audience, not all of those details necessarily need to be in the communication. Right? In most cases, your audience is trusting that you’ve done that robust stuff behind the scenes. And so then think about, whether and how much detail needs to be in the communication piece of that. So maybe my first tip would be to simplify.
Cole: 00:24:26
Second would be focus attention. So be clear about where you want your audience to look in what you’re showing and take steps to focus their attention there. Often the easiest way to do so is through color. If you do things in gray scale or black and white and then use color sparingly and strategically just on the most important thing, it directs attention there very quickly, but any sort of contrast will have this effect when it’s used sparingly. So that would be my second tip, focus attention where you want people to look. And the third tip would be your use of words, use words. I think sometimes when we think about visualizing data, there’s this misconception that it should all be numbers and pictures, but we need words to make those numbers and pictures accessible to people. So there’s some words that have to be there, and every graph should have a descriptive title. Every axis should have a title, even if you think it’s totally obvious, just the lack of these things can cause there to be some questions on the part of your audience.
Cole: 00:25:29
Or they may have to make an assumption and you don’t want them to use their brain power to make assumptions, just title explicitly then they can use their brain power to focus on the data and the message. And then pairing these last two, right? Focus and words, show someone where you want them to look and then put words on the page or in your voice over, if you’re talking through something live that tell your audience why you want them to look there. I think another common misperception is that we should allow the data to speak for itself and data can, the challenge is just that it has the risk, it runs the risk of saying something different to every different person who’s looking at it. And so if you have a message that you want people to get out of the data or something specific you want them to know, direct their attention there and put that into words so that it’s clear.
Cole: 00:26:18
And those words actually do some really interesting things psychologically for the framing of when we look at data. Studies have shown that, for example, if you title a page or a section with the takeaway that you want people to see in the data, and then you show them the data, they’re more likely not only to focus on that takeaway and the data, but to remember that takeaway after the data is no longer in front of them. So choosing titles and the words that you put around your data is a very powerful thing and you want to do that carefully. Those would be my three tips, simplify, focus attention, and use words wisely.
Kirill: 00:26:55
Love it. I think often people are, when people struggle with the challenge that they have so much interesting stuff to present in one chart and they’re like, okay, I’ll put this here. And then this age group is doing this, and this age group is doing that. And these customers are doing that or whatever else. And like, you want to share all 10 insights that you got, but you end up sharing zero. So I guess, following your advice, it’s better to share two that people will remember than trying to share more, less is more in this case.
Cole: 00:27:26
Well, and this is where story can come in, in really helpful ways because when you’re analyzing the data, you’re exploring the data, you need to look at all those 10 things, right? And more, it just doesn’t mean that we need to communicate all of those things necessarily. And this will always be context and audience dependent. But for any analysis, when we’re exploring, we’re trying to understand where there are interesting things that we can learn from that might impact how we do things or drive us to make smarter decisions, you need to look in a lot of different places, right? To get a robust picture of that. But then at the end of that, you want to step back from the data, step back from the analysis and think about, okay, now given that, given what I’ve learned, what does my audience need to know? And they don’t need every single bit of data you looked at.
Cole: 00:28:09
Again, they’re assuming for the most part that you’ve done a robust analysis, that is your job, your job is to tell them what do they need to do now. What are the important things to know now that you’ve done this robust analysis? And it’s not, I looked at these 50 things. It’s I looked at these 50 things, don’t even worry about that. We can get into that detail if you want to, but here are the two things upfront, right? Simplifying, focusing, using words to make the point clear that you need to focus on today or in the near term. And I think too often, people who are in analytical roles, who are doing a lot of analysis of data, they want to simply present the data, right? And assume that the audience will know what to do with it.
Cole: 00:28:54
They’ll have context, they’ll treat it appropriately. But the challenge is audiences are faced with just increasing amounts of data all of the time. And so when we simply serve up more data, it’s really easy for an audience to say, oh, that’s interesting. And then they turn their attention onto the next. Or worse, they ask you for even more data, right? It’s going to become this sort of death by data that I mentioned earlier. Whereas if you take it to the next step, you say not only audience here’s the data, but here’s something you should do with this data. Here’s a discussion to have, or a decision to make, options to consider, even if we don’t have the full picture and we recommend something that might not be the right path, it gets your audience thinking about action and focused on the right sort of conversation. And that’s a conversation that often gets skipped when we stop at simply showing data.
Cole: 00:29:46
And so I’m a strong advocate for anyone who’s analyzing data and using it to communicate, put a stake in the ground. Have an opinion on what now should happen next and make that clear to those, to whom you’re communicating because that is how we can derive more value from the work that you’re already doing. Because I am a huge believer that there is incredible amount of value to be obtained from great work that’s already being done, that just isn’t being communicated as effectively as it could be. So really thinking it through leaving time for this part of the process, right? Because the communication piece, but whatever reason, I don’t know if it just, it seems less technical maybe or less sexy. I don’t know. We tend to overlook this piece, right? You spend all this time doing a fantastic study or analysis and often it’s time constraints too, that like we get done. Okay. We’ve plotted a graph, we maybe put some words around it and we’re done.
Cole: 00:30:43
But that graph is the only part of any of that work that anybody else actually sees. And that is the moment when things either succeed or fail. So the communication of your analysis is as important as every other step in the process. And so I think helping people recognize that and spend more time on that, right? Practicing, figuring out what works and what doesn’t and getting back to the sort of things that is covered in the books. And actually, this is maybe a good segue into the Storytelling With Data community, because I am a strong believer that to get really good in this space means practicing and getting feedback and talking with other people who may be encountering similar challenges or have come up with creative solutions.
Cole: 00:31:32
And this stems back to the fear that I mentioned earlier of somebody who goes to a workshop or they read the book and they think yeah, that’s great. But then get sucked back into the day to day of normalcy and nothing actually changes. And so giving people a low risk spot to practice more and get feedback. And so the Storytelling With Data community is an online community that has been set up to facilitate all of these things. So it’s free, it takes a couple minutes to sign up, but you can go there. This is at community.storytellingwithdata.com and there you’ll find ways to get feedback, that you can post anonymized examples from your work to get feedback or things you might be trying out. You can have conversations with other people. And then we have ways, low risk ways that people can practice.
Cole: 00:32:17
So for example, we do a monthly challenge that we’ve been doing for coming up on three years now. So every month there’s a new challenge and this ranges from a specific graph type. So for example, create a line graph and you’re meant to go out, find some data of interest. We list a couple of hundred public data sources in case you’re looking for data and build like your best line graph. And so it gives people a reason to look for something interesting that they may want to visualize and practice, and then you submit it in the community and then other people can comment and give feedback, or you can ask for specific feedback. And so it ends up being great, both from the standpoint of giving the folks who undertake it a reason to practice something specific. But then also at the end of it, it becomes this fantastic archive of, at the end of that, we have now 88 annotated line graphs, for example, that now people can go back and look at for inspiration or to figure out what works better or worse.
Cole: 00:33:11
And so the current one that we’re running right now is designing a diagram, which is sort of a little bit out of the data viz space, but still in this visual sort of realm of take a process or a concept and make a diagram. And so you can see there are a couple of dozen people already who’ve contributed this month. And then we have an exercise bank. Then the exercises are very similar to what you’d find in Let’s Practice, where everything is provided. The data’s provided, the scenarios provided. And you’re asked to flex a really specific skill, whether it be using color and words like we talked about earlier or some other best practices. And then there are a few that are more full blown case study, practice at all sort of exercises. And so we’ve got, I don’t know, probably 30 or 40 exercises there.
Cole: 00:33:58
Now we’re adding to that exercise bank all the time, so you can filter to be able to practice specific skills. And then we just, about two months ago introduced premium in the community, so everybody, everything I’ve talked about so far is free. But then there’s another level that is a subscription model where folks who sign up for premium get access to, we do monthly live events where I or the team talks through a lesson, maybe does a make over and people can join live and contribute questions and ask, and get some insight from the team that way. We do regular office hours and then we have library of on demand video learning as well.
Kirill: 00:34:39
Wow, sounds like you’ve built a very big ecosystem around data visualization and it’s inspiring because as you said, just before your book, the only places to go to where the Tufte’s principle maybe some more kind of like Cleveland and McGill ranking of elementary perceptual tasks, like from the 90s or 80s, is great to see that there is a community building coming around here and awesome that people can share. And we actually also have a community and they’ve already submitted some questions for you, which I’d love to go over a bit later on this podcast. But before we jump into that one more very important topic that I don’t think I’ve covered off in depth with any guests on the podcast previously, and I think you’re the perfect guest to cover off with is storytelling. Can you please tell us like what is a good story? What goes into a good story? How does one structure a good story?
Cole: 00:35:40
Yeah, great question. And there are multiple ways to structure a story, right? There’s not one right way, but when I’m thinking about story and when we’re teaching about story, we usually do so through the narrative arc, where if you picture an arc, you start off with the plot, what’s happening at the beginning of the story, then some sort of tension gets introduced. That tension builds in the form of a rising action. It reaches a point of climax. Then there’s a falling action and finally a resolution. And so we teach this and we show it in this arc sort of shape that does a couple of things, because I think oftentimes when we’re in a technical role, when we’re analyzing data, often the way we communicate that ends up mirroring the process we went through.
Cole: 00:36:24
So it’s a more linear path where we might start off with what’s the question that we set out to answer in the first place. What was our hypothesis at the onset, then there’s the data, right? Where did we get it? What assumptions did we make? Did we do things to clean it? Then there’s the actual analysis we went through, right? What was the statistical methodologies that we employed and then our findings. And so it’s this very linear path. And the challenge with the typical linear path is that it’s very selfish in terms of, it comes most naturally because that’s the path I went through when I was analyzing the data. But we can do the linear path without ever pausing and really considering our audience. And for me, that’s the biggest shift that happens when we think about not communicating in a linear way, but rethinking our analytical communications along the path of a story.
Cole: 00:37:09
Because to have a story, you have to have tension and it’s not about making up tension, right? If there weren’t any tension, you’d have nothing to communicate about in the first place. It’s about figuring out what that tension is, how do you bring attention to it and it’s not the tension that matters to you typically, it’s the tension that matters to your audience, which means we have to do this sort of mind bending. And instead of communicating for ourselves, our own needs, for our data, for our analysis, communicate first and foremost for our audience. Then when we think about story structure, the resolution at the end becomes what your audience can do to resolve the tension that you’ve brought to light.
Cole: 00:37:52
And what happens is when you communicate this way, thinking about your audience, structuring things for them, drawing their attention to the tension that’s going to matter to them, that also matters to you, but making those things work in coordination with each other, you can get people to pay attention in new ways. You can motivate them to act in totally different ways that is really powerful. So I think it’s interesting. I sometimes get the reaction and more often from technical folks in a story that’s like fluffy marketing stuff. You can actually be very strategic in the way that you use story to communicate. So the narrative arc is one way of thinking about that.
Cole: 00:38:30
And what we’ll often do in practice with a group is everybody has little sticky notes, write little posters, and at the onset, we’ll have them get a project in mind and just start story boarding, just start writing down pieces of content that maybe will fit into the eventual picture. And at first, just brainstorm. Don’t worry about order. Don’t worry about what makes it into the final deck or report, just brainstorm. And then you start organizing those post it notes and figuring out how do you group them together, right? What order might you want to go through them? And often at the end of that process, we end up with that sort of linear path that I described before. And then we think about story, okay? Now that we’ve got these pieces, how might we look at this through a different lens and take our audience along an intentional path that will work for them?
Cole: 00:39:19
And there are a ton of ways that you could do that, right? And that’s where you can try out different things or take what you know about an audience. In some cases it might make sense to start with the finding. Hey, we did this analysis and we found X. In other cases, you’re going to need to build up to that cause if you start there and particularly if that runs counter to people’s expectations at the onset, it may go down a bad path. And so maybe there’s some cases where you build up to that or where you take people through different paths to get there. But so the main thing is being intentional about how you communicate, because I think too often we fall into this trap where we do it the same way we’ve always done it because we’ve always done it that way.
Cole: 00:39:59
Whereas when we’re thinking about who our audience is, how we’re delivering the information to them, what biases they have, what constraints we’re facing and we do things to try to make all of those inputs work together, that we can align ourselves to have more success when it comes to how we communicate. In stories a piece of that, I think the overall puzzle.
Kirill: 00:40:23
I love it. I love it. How many steps do you recommend for a story, is it five, 12, 20, what’s the optimal number?
Cole: 00:40:29
I don’t know that there’s an optimal number. I don’t think there is an optimal number. There’s probably some sort of minimum that you wouldn’t go below, I’d even question that. I think it’s going to depend, right? It’s going to depend on what level of detail your audience needs. This is another thing, right? For the same analysis for different audiences, you may take them through different stories. So if I’m communicating to the C-suite, it may be a three-step story, right? Or maybe they actually just care about the, so what, and they don’t need any of the other stuff. Versus if I’m communicating to, just making things up, like my finance partner, and they want all of the detail, they want to know cost associated with different things or … I’m just making things up, but where the level of detail may vary quite a lot depending on who the audience is.
Cole: 00:41:15
And this is another thing to be thoughtful about is do we need to communicate to everybody all at once or should we communicate to different audiences separately? And there are pros and cons to both of those things, but I’m always an advocate of, if it’s something really important and your audience’s needs are sufficiently different than trying to communicate to them differently so that you can match your given audience’s needs each time.
Kirill: 00:41:37
Got you. And how do you see the difference between communicating a presentation in person where you can actually present in front of a projector or virtually versus sending somebody just the PowerPoint slides and they have to go through it? Like, do you have different approaches to that?
Cole: 00:41:54
Yeah, absolutely. When you’re there in person, whether it’s in a room with somebody, or in our two dimensional environment of today in the virtual world, your materials can be relatively sparse because you’re there to talk to them and you can speed up or slow down and pause and answer questions, or go into more detail or less detail depending on what people need. And actually one strategy that works very well in person, but especially in the virtual environment, and particularly if you’re going to, if you need to show something complicated is build it piece by piece, right? So if I want to lead up to a graph that’s going to look kind of gnarly at the end of it, but I need to do that for some reason, I don’t need to start with the gnarly looking graph. I can start with just the axes, right? Put up the framework, the scaffolding.
Cole: 00:42:39
Then I can show just the titles and labels on those axes, which forces people to sit with me through the explanation of what they’re going to be looking at before they jump to the data. Because I haven’t even shown them the data yet. And then you layer on the data sometimes piece by piece, or if there’s a chronological component, you can walk them through how something builds over time. And so you build up to something that at the end might look complicated, but because you’ve broken it into these piece-wise steps, it doesn’t feel complicated anymore. And this can work great in the virtual environment because it keeps people’s attention on your slides or on your screen because of that motion. Right? And it builds this bit of anticipation of people will be less likely to multitask over in their email because they might miss what happens next.
Cole: 00:43:23
And it builds in some attention that can be useful in the virtual environment, especially. So slides, communications can be sparse when you’re talking through them because you can land a lot of that detail through your voiceover and you don’t want the visual competing with attention for what you’re saying. Versus when you send something around the level of detail that needs to be physically written in that document is much higher. And our tolerance for detail is higher in that setting, right? When I’m reviewing something just myself on my own computer, I will read through things more than I might be apt to if someone flashes a slide in front of me in a live presentation. So these two scenarios for how you’re delivering data, the biggest thing that changes is the level of detail, much more in that thing that you’re sending around because you’re not there to lend the voiceover and answer questions and lend context, which mean all of that has to be done on the page itself.
Kirill: 00:44:18
Got you. And you teach all these things in the workshops that you hold-
Cole: 00:44:22
We do.
Kirill: 00:44:23
And you have a workshop coming up on the 15th of September, a public workshop. Tell us a bit about that.
Cole: 00:44:30
Yeah. So, these have gone virtual as well, and we do these periodically, usually a few a year in different cities around the world, but we’ve gone virtual in 2020. And so our next one will be September 15th, all that information’s on the site at storytellingwithdata.com/public-workshops. And it’s similar to what I was describing before with some of the interactivity that we’ve been using, where this is a five hour session, draws people from all around the world, all different sorts of organizations and industries. And so part of the fun is just the variety of attendees and you’ll have ways to interact with other attendees, but it’s a mix of teaching some foundational principles for communicating with data, and then a lot of hands on practice where there’s exercises. I know it’s almost similar to the pairing of the first book and the second book, it’s a little bit of teaching, a lot of hands on practice and then integrating interactivity.
Cole: 00:45:25
So both interactivity with the folks facilitating, which is me and the team, and then also interactivity with the other participants. And it’s super fun, right? Five hours sounds at the onset, like a long time to sit in front of your computer, but the time flies by, and actually the biggest piece of feedback we got from the last time we did this in May was that people wanted more, they wished it was longer. Like, I don’t know, five hours is already pretty long to sit in front of a computer, but we mix things up in ways to keep it really engaging and interactive. So I’d say for anyone who wants guidance in honing the kinds of skills that we’ve talked about today, it’s an excellent place to do so.
Kirill: 00:46:03
Fantastic. Thank you. Thank you very much. It sounds very exciting. Do you have preferred tools, like for instance, in this workshop, do we have to use Tableau or Qlick or is it like, can I use any tool?
Cole: 00:46:17
Great question. So we keep the workshops low tech, that we recommend tools like post it notes like I mentioned before, pen, paper, where you’re sketching ideas and there’s a couple of benefits to that. One, it’s faster, right? So we can iterate and have things to show other people in a low tech way without getting bogged down by tools and the time that that takes. But also it frees us up from the constraints of our tools or what we know how to do in our tools and just get the creative juices flowing in ways that’s useful as we’re thinking about how we show our data in new ways. And then when it comes to the tools that we use primarily, if you look to the books or the examples on the site are primarily Excel, primarily Excel and PowerPoint.
Cole: 00:47:01
We do some Tableau, but we end up going back to Excel primarily just because it’s so universal, right? Everybody has it. Everybody can make graphs in it. The learning curve is relatively low. And I love the fact that anyone can pick it up and make a graph. The challenge is just that in most cases nobody’s really taught us how to do this. And it’s easy to do some bad things, but for me, the tool, the lessons we teach are tool agnostic. It can be done in any tool, right? And you’ll face different constraints in any tool, but that’s part of the fun, right?
Kirill: 00:47:34
Awesome.
Cole: 00:47:34
It’s filtering out what are the constraints I face this time and how do I still make it work? And that’s where getting a piece of paper and getting it right on paper first, and then looking to your tools to try to make it real can be a good way to go.
Kirill: 00:47:46
And I love that on your website and you share some links in your book as well. You share these templates for Excel, where somebody can just like download a template and build a complex chart which is cool.
Cole: 00:47:59
Yeah. And actually I should mention, so all of the data, all of the graphs for Let’s Practice are available in the first book too, but for the second book, we’ve really been concentrating on getting solutions built in different tools. So all the Excel solutions are available. And then there are some exercises that we’ve built solutions and tools ranging from Google data studio to R, Python, Tableau, Power BI. And there’s actually, there’s a community member, Adam Roboto. I might be saying his last name wrong, who just posted. He actually went through the first book and recreated all of the graphs in R and made his library of code available in GitHub.
Kirill: 00:48:36
Amazing.
Cole: 00:48:37
And so if you go to the site, the storytellingwithdata.com/book/downloads, or /letspracticedownloads, you’ll find everything that we know about, that’s been done in these other tools, because it’s a fantastic way to learn, right? Is you take something that you can see and now try to recreate it in your tools or get as close to that as you can.
Kirill: 00:48:55
Fantastic. Love it. All right, Cole are you ready for rapid fire questions from our community?
Cole: 00:49:02
Oh, sure, let’s do it.
Kirill: 00:49:03
Okay, here we go. So I posted on LinkedIn, lots of people have replied. And here we are going to start with a question from Julian who is asking, “What are Cole’s tips for determining the right level of simplicity and amount of information of the visualizations for the target audience?”
Cole: 00:49:22
This is a hard question, Julian, hard to answer specifically, right? When we talked about simplifying and focusing attention earlier. And so really think about your audience, because one of the things that we have found in practice is if you only simplify, right? If you focus on decluttering, but do nothing else, it can leave your audience feeling like you’ve taken things away without adding back the requisite value. And so if you’re taking the time to declutter also make sure that you’re taking the time to focus attention and highlight the takeaway, because then that more than what makes up for the things that have been taken away. But when it comes to the right amount of detail for a given audience, use what you know about them, about how they like to be communicated to, what level of detail they’re going to desire.
Cole: 00:50:09
And if you don’t know them directly, do you know colleagues who’ve communicated with them or do you know people like them? And sometimes you have to make assumptions and try to go in and learn from each scenario of, okay, that was too much. We got down this path, I didn’t mean to, next time let’s scale it back. And that was not enough, we didn’t get into the media discussions. And so it’s a little bit of trial and error, but then of course, learning from each experience and learning from your colleagues around you, who may have insight.
Kirill: 00:50:33
Got you. Awesome. Next question is from Bruno, what are best practices to consider when designing a mobile dashboard?
Cole: 00:50:41
So I don’t spend a lot of time in mobile, but we actually did one of our challenges, it was a guest one that Andy Cotgreave did for us. He’s one of the authors of The Big Book of Dashboards. This was last summer, so it’s probably last July or August, maybe last August but the challenge was designing for mobile. And so if you go to the challenge archives in the community, you can see a lot of different approaches that people took. And he mentions a ton of resources, I think, in his writeup as well. So that would be a place to check out, but you want to think about how people are going to be interacting, right? They’re on a much smaller object than a computer, and so the way you scale things is differently, how you arrange things. There are considerations of are people swiping or are they scrolling? And this is not an answer to the question, but some of the things to be thinking about.
Kirill: 00:51:31
Fantastic, and we’ll link to that resource in the show notes if anybody wants to check it out. Maximilian asks, what are the three most often used visualization types in your daily work?
Cole: 00:51:42
So we spend most of our space in business communications, right? Where it’s, something’s being communicated in an organizational setting. And so the charts that we see most commonly are going to be the familiar ones, they’re line graphs and bar charts. And what would the third be, maybe a pie. I don’t know, just a little bit on that, and there’s good reason for that, right? Because people already know how to read these graphs, and so you face less of a learning curve for getting your information across. So you always want to think about if you’re choosing something less familiar or more novel, one, be thoughtful about why you’re doing that. And secondly, recognize that you’re introducing a hurdle, you’re going to need to get your audience over that hurdle and understanding how to interpret the visual before you can even start talking about the data.
Cole: 00:52:31
And so it means oftentimes when we’re analyzing data, we may be using very different visuals compared to when we now need to communicate that data to someone else. And when it comes to the communication piece and again, the audience context depend, but oftentimes it is the simple, straightforward graphs that are going to do that most effective.
Kirill: 00:52:48
So sometimes you want to get creative and like do something amazing, but oftentimes stick with the simple stuff to get your point across.
Cole: 00:52:57
Yeah. You don’t want it to be at the expense of your message being heard or understood.
Kirill: 00:53:02
Yeah. I guess there’s like two styles of visualization, and one for work, and one, if you want to go publish your dashboard, impress others, or just like have fun, you can share that in for instance, your community, right? And just explore new skills, but you should separate the two, there’s work and there’s play.
Cole: 00:53:19
Well, and there are different reasons for visualizing data, right? In some cases we visualize it to understand it better, in some cases we visualize it to communicate something to someone else, in other cases yeah, we might just be having fun or wanting to create something beautiful. But to your point, you just want to be thoughtful about when you’re optimizing for which of those things.
Kirill: 00:53:36
Got you. Another question from Maximilian, what defines an excellent one slide data visualization from your point of view?
Cole: 00:53:46
From my point of view, it would draw on a lot of the things that we’ve talked about already. So for me, an excellent data visualization, when it’s put in front of me, I know where to look, everything is titled and labeled in a way that if I have questions about what I’m looking at, I can find the answers to those questions. I’m not finding myself having to make assumptions or not really sure what I’m looking at. And for me, the aesthetically pleasing bit of it is important as well of, I want it to be something that I want to look at, right? Where someone was thoughtful in how it’s designed aesthetically, that it’s free of unnecessary clutter or things that don’t need to call attention are pushed to the background, right? That there was thoughtfulness in the design, I guess, is how I’d characterize that piece.
Kirill: 00:54:38
Okay. I think that’s a very valid point. And we did speak a lot about these things already. Bernardus asks, what are the best practices to teach non data people for quickly grasping the concept of data visualization, especially for explanations? Something, I guess that’s dear to your heart because you teach a lot of people. So what are your best practices for teaching people?
Cole: 00:54:55
This is a great question. And this is a hard question to answer in a rapid fire sort of situation, right? Because for me, and there’s a whole topic there. There are books written about data literacy and how do we improve people’s data literacy, but I’d actually take it I think a step back from that to say, don’t focus on the graph, right? The purpose is not to teach someone how to understand the data visualization in most cases, the purpose is to get them to understand what you need them to know, and you can use graphs to help do that. And I think when you employ some of the strategies that we’ve talked about in terms of don’t over complicate, use a straight forward sort of visual that people are going to, that’ll be familiar, that they’ve seen before.
Cole: 00:55:43
So you’re not overwhelming someone who says, ah, I’m not good at graphs, I’m not going to get this. Where it’s something straight forward, so you’re not scaring people. Then it’s really, it’s the story. It’s the narrative that you build around it because when that’s done well, when I have a compelling narrative, I’ve made it clear to my audience where they’re meant to look and what I’m showing, whether that’s the words or the data or combination of those things, I’ve thoughtfully gotten rid of extraneous things that don’t need to be there, then your audience is focused on the ‘so what’ almost to the extent that they may not even realize they’re looking at a graph. And there will be cases where no, it’s actually, it’s this really nuanced visual we need to use because it allows us to have this insight that otherwise we can’t see.
Cole: 00:56:28
And so there will be cases where you need to teach your audience how to look at something or what to see or how to do that. And in those cases when you can, sit side by side so that you can, or maybe not side by side today, but sit one on one with that person and talk them through piece by piece. You can build in the way that we talked about a little bit earlier, if that makes sense, but have them ask questions and get an understanding because I think it’s too easy for people when we put a graph in front of somebody else and they don’t get it to say, oh, that’s their fault. But that’s a failure on the part of the design, right? If we put a graph in front of someone else and they don’t get it, that is our fault as the designers of that graph, it means we did something that made it unclear or made it impenetrable in some way.
Cole: 00:57:11
And so learning from that of what could we do or when you can ask the question, what’s confusing here, or what caused you to focus on this one weird thing that I didn’t think anyone was going to look at and iterate from that. And a great way to do that before putting in front of your audience is just iterate with people around you, right? Colleagues, friends, and particularly people who are much less close to your work than you are. Put something in front of someone and have them talk you through what they see, where they pay attention, what takeaways they might highlight from it and learn from this process. Because as we get close to our data, it becomes really hard to have fresh eyes when we’re looking at things, so using other people and getting their fresh lenses. And if something’s not sitting right with the audience, don’t assume the problem sits with them. Assume the problem sits with you and the design and work to better understand what’s causing any resistance and how you might get at that more directly.
Kirill: 00:58:13
Amazing. I think people should embrace and celebrate those moments when somebody says I don’t get it, because that means you’re about to step outside your comfort zone and learn something new. How to explain better. The easier way out is to blame the person, but as you said, take responsibility, take ownership and you will grow and learn as a designer in the process.
Cole: 00:58:31
Yeah, and I love that. It’s fantastic when someone says I don’t get it because then you can have a conversation about, well, what’s challenging, right? Tell me more about that, because I think more often people don’t admit it and you just, you’ve lost the opportunity to communicate because now somebody doesn’t get it, but they don’t voice that and you don’t know, so you don’t ever get to have that conversation.
Kirill: 00:58:49
Yeah. Got you. Okay, next question from Kyle, “Really excited for this episode, I was just curious how the role of psychology and visual perception plays in Cole’s design process and what are some important principles that she always tries to keep in mind?”
Cole: 00:59:10
Good questions. So I think when we think of the psychology, I go mostly to how people’s brains work and where they’re going to look and what sort of things will draw attention. So the things that I like to keep in mind when I’m designing my own graphs contrast is a really big one. And I talked about that a little bit earlier with color, but really employing contrast thoughtfully. And you can do that in a lot of different ways. You can do that by sizing things differently or sparing use of color, like we talked about, but this draws on the pre-attentive attributes, right? Both books go into those in more detail, but things like color, orientation, shape and such that when everything’s the same and you vary one or a couple of these in ways that are really contrasting, that draws people’s attention.
Cole: 00:59:59
And so I’m always thinking about where do I want people to look, or how do I want them to interact with the information, right? What’s the visual hierarchy? Where should they look first? Where should they look next? And how do you orient things on the page and use contrast to make that clear. Where sometimes there are less important things or like it’s like the third or fourth order of importance, but things that still need to be there where you can push them to the background and make them gray, use smaller texts, do things to de-emphasize so that you create this really clear visual hierarchy. And so I’m always thinking about that. And I’m also always thinking about, and this came up before, but how I’m using my words. Using accessible language, making sure things are titled and labeled because the things that are clear to me about my data, because I know my data aren’t necessarily going to be clear to somebody else looking at my data.
Cole: 01:00:49
So really taking intentional steps to make it easy for my audience to know where to look and what to see. And then getting feedback and especially getting feedback if you’re trying new things or doing something in a different way, because we sort of … And when we try something new, we can fall in love with the process, right? Where it looks great to us because it’s this new shiny sort of thing, whether it’s, I’ve stripped out color and I’m using color only in a certain way now, or I’m trying out a graph type that is maybe less common that you can get really attached to that. And particularly if it’s taking you a long time to generate it, right? That we form this attachment. So getting feedback and letting go of things, even if you’ve invested time and thought of, if it’s not working, letting yourself iterate and let go of things.
Cole: 01:01:40
And that’s one of the reasons that starting with paper can be so important because you can quickly sketch something crazy on paper and realize, oh, I thought that would work but it actually doesn’t and just recycle the paper. But if I’ve just spent hours of my day or days of my week creating this thing, it’s much harder to let go of. And so being thoughtful about how we iterate and looking at data a ton of different ways can be really insightful, both in terms of figuring out how might we communicate it in a way that’s going to work for someone else, but also just to help us get to know our data better. There are different nuances we can see when we cycle through, right? We try it in bars, we try it as lines. We throw it into a scatterplot. We do these different things because it will allow us to get to know our data better so that then we can talk about it more fluently and can land on these ways to show it that’ll help our audience understand it more quickly.
Kirill: 01:02:31
Wow, thank you. Amazing answer. You covered so many great points. Exhaustive answer as well, nothing to add.
Cole: 01:02:39
Well, you said lightning round and I didn’t warn you that I can talk for a long time on these topics.
Kirill: 01:02:45
No, love it. Love it. One question left, this one might be tricky, I guess saved the best for last unintentionally, it just happened, but, let’s see. Let’s see how we go. So Monica asks, “Has there been a time when you were creating a visual for an audience you do not know well (I.e. for example, a new job or a contest)? Since knowing the audience is a big part of creating a useful visual, what are some tips?”
Cole: 01:03:10
Yeah. Great question. And so, for audiences that you don’t know, or you can’t talk to directly, then you want to think about, well, what assumptions can I make about them? Right? What do I know about them? What do I know about what they might care about? Or what could I assume about what they might care about? And so you can oftentimes make assumptions, right? It would be one place to start, but test those assumptions. Think about, well, what if I’m wrong here? And this is where getting colleagues involved and brainstorming can be useful, but it’s not necessarily that you have to know your specific audience, right? Because you may know other people who are similar to that audience or who will have biases that are like that audience, or you may have colleagues who have communicated successfully or unsuccessfully with a similar audience before, and so trying to learn from that process.
Cole: 01:04:00
And I don’t remember which one offhand, but I know there is an exercise in Let’s Practice. It would be in chapter one, which is about context and thinking about our audience and our message, that is some thought starters for getting to know your audience. And there are some specifics for when it’s an unknown audience of things that you can do there. But I think any time you’re spending thinking critically about who your audience is and what they care about, even if you’re making some assumptions and even if some of those assumptions are wrong, that is still going to be time well spent, because now you’ve gotten out of your own head and started thinking about things from someone else’s point of view. And that is always going to be helpful when we communicate, even if we get the details a little bit wrong and then learn from each time, right? Of what worked well, what didn’t work well, do thoughtful postmortems, so that you can iterate and continue to hone how you’re doing things over time.
Kirill: 01:04:51
And that will probably also help create a kind of like these archetypes of audiences. And then next time it’ll help you be like, oh, I have five archetypes in my mind, or even written down somewhere, which archetype does this audience fall under? And then you already know what to do. And if it doesn’t fall under any architect, create a new one, the sixth one, the seventh one, so on.
Cole: 01:05:09
Yep, absolutely.
Kirill: 01:05:10
Great. Awesome. Cole, just one more question, I guess from me, to wrap things up. I love to ask this question. It’s more of a philosophical type of question, from what you’ve seen in the space of data visualization, in the world and how it’s been changing, and you’ve been in this space through many different, the good and the bad times. And you’ve been at Google and running your own business and helping lots of organizations. Where do you think this whole space of data visualization is going? What would it look like in three to five years? What should our listeners prepare for if they want to build a career in this space?
Cole: 01:05:52
Yeah. Great question. And I think one of the things that has happened over the years is that jobs focused on data used to be a sort of niche thing and that’s no longer the case, right? We have data all around us. We’re capturing it everywhere and so increasingly doing things with data has become parts of people’s roles that never historically had to deal with data. Right? I think of just the Google example, human resources, not a data driven organization historically. And so now you have people in like HR generalist sort of jobs, who are having to look at data and having to interpret and do things with it. And that’s happening at scale everywhere, right? Where roles that historically didn’t need to know how to do things with data are increasingly needing to do so. And so the skill that comes with that, that is in desperate need is how do you communicate that data effectively?
Cole: 01:06:48
So I, for your audience here would be a huge advocate of any time you spend investing in yourself and how you can communicate your data, both how you show it, right? Some of the visual components that we’ve talked about today, but then also you and how you talk through your data, how you present it becomes such an important piece because … And those people who are good at that, or get good at that are going to be in high demand because I think it’s interesting, there’s been in the last few years, what feels like a big emphasis on data science, right? It’s just like sexy new role. Like yeah, everybody wants to be a data scientist. Like that sounds really cool. It existed for a long time, they just used to have names like statistician, right? Which does not sound as cool. But there’s been more awareness around an investment in skills related to data science. Right?
Cole: 01:07:42
Which is hugely important, but has been a little bit at the expense, I think. And I think where we’ll shift next is now this other piece of, hey, you can analyze the data, right? You can pull insights from it, but how do you verbalize those to somebody else in a way that they can understand and take action on it? So any investing you do in your skills, when it comes to how you’re communicating your data will serve you very well I think in the coming years, because I think that’s what we need, right? So that we don’t get overwhelmed by data, but we have people who can thoughtfully turn that data into information and that the world can do greater things as a result of that.
Kirill: 01:08:21
Absolutely. And to add to that, you at the beginning of the podcast, you mentioned that presenting data is as important as building the models. I would say it’s more important and the reason for that is like for instance, as a business owner, would I rather somebody build me a cool, super high tech, deep learning XG boost model and not be able to explain it to me and the insights, or would I be satisfied with a logistic regression simpler model, but that is presented in a way that I know exactly what to do? Seven days out of the week I will go for the second option. Data presentation, in my view, number one skill for all kinds of data scientists.
Cole: 01:09:10
I concur.
Kirill: 01:09:11
Awesome. Cole, thank you so much for coming, it’s been a huge pleasure. Before I let you go, where can our audience find you, follow you, get in touch, get to know about your books and all these amazing things that you’re doing?
Cole: 01:09:23
So you can find everything that we’ve talked about today on our website, which is storytellingwithdata.com. I’m also active on Twitter, I’m @StoryWithData. We have a LinkedIn page for Storytelling With Data where we post tips and articles and resources on a daily basis, but start with the website, storytellingwithdata.com and you’ll find everything else from there.
Kirill: 01:09:44
Fantastic. Love it. And of course, I encourage people to follow you on LinkedIn. I think you have 10,000 followers there who are benefiting from all the things you’re sharing. And I love the idea of your community. So people should join that.
Cole: 01:10:00
Yeah, that’s fun.
Kirill: 01:10:00 And I wanted to say, I totally love your book. If anybody’s reading this book right now, we even have people who posted in their questions, like I’m reading the book now, I’m finishing it up. The previous guests on the podcast, he was just talking about your book. If you’ve read the book, leave a review, it is available on Amazon, live a review on Amazon, it’s totally worth it, so more people can learn about it.
Cole: 01:10:20
Thank you. That’s fantastic.
Kirill: 01:10:22
Awesome. Well, thank you very much Cole, it has been a pleasure chatting with you today.
Cole: 01:10:25
Thanks for having me Kirill.
Kirill: 01:10:32
So there you have it. I hope you enjoyed this podcast. Thank you for sharing this hour with us. And what was your favorite takeaway? My personal favorite takeaway was the idea that Cole shared about building a complex visualization. So of course we have to simplify and make it easier, but if it has to be complex, do it step by step, do the axes first and add some elements, then some other elements, maybe some motion so that people see how the visualization is being built up. And of course there is so much value on this podcast, I hope you were able to get as much as you can, and you’ll be able to implement these things in your career because data visualization is one of the most important steps in a data science project.
Kirill: 01:11:18
As always, you can get all the show notes for this episode at www.superdatascience.com/395, that’s www.superdatascience.com/395, there you’ll find the link to Cole’s website, the community, her LinkedIn profile, her Twitter and everything else, including the workshop that is coming up on the 15th of September. Now about the workshop, first thing is that Cole created a coupon for our audience. So if you want to attend the workshop, make sure to put in the coupon SUPERDATASCIENCE, all one word, at checkout and you’ll get 10% off on this workshop. Sounds like an amazing workshop. I’m personally right now considering of attending and I want to see if I can fit it into my schedule. But if I can, I would love to sit through this five hour workshops. So if you are going, they have coupon code for podcast listeners is SuperDataScience to get your 10% off.
Kirill: 01:12:13
Also SuperDataScience will be sponsoring one of our listeners to go to this workshop completely on us and that could be you. And the way to participate is leave a review for this podcast, just go to Apple, iTunes or wherever you’re listening to this podcast, leave a review, take a screenshot and send it to podcast@superdatascience.com and we’ll pick a random winner from everybody who sends a review. It doesn’t have to be a five star review, if you think it’s a terrible podcast, leave a terrible review, but take a screenshot, send it to podcast@superdatascience.com and we will pick a random person, absolutely random from everybody who sends them in and you will get a receipt out to this amazing workshop.
Kirill: 01:12:57
So in any case, if you want to attend, go and get your seat now, and then submit your review and maybe you’ll get another seat and then you can share it with a friend or colleague or family member, share the love, spread love. So you go, very simple, all the show notes are at www.superdatascience.com/395 and make sure to leave a review, send it to podcast@superdatascience.com. Hope you enjoyed this episode, I look forward to seeing you back here next time, and until then, happy analyzing.
