When you need to share data with an audience—like in a fundraising presentation or a project proposal—it’s tempting to just toss together a graph or chart, tack on a title, and call it a day. But are you sure your message is really getting through?
In Episode 3 of The Analyst, she shares:
- Why data storytelling matters, for everyone
- How to filter out noise and focus only on the information that matters
- Tips for selecting graph/chart types, writing copy for your visualizations, and more
Sarah Siwak: In a couple of sentences could you tell us what you do?
Cole Nussbaumer Knaflic: I teach people how to tell stories with data. This takes a couple of different forms. I spend most of my time teaching workshops for organizations, where I go in and spend half a day or a day with the team going over foundational lessons for telling effective stories with data, also looking at some of the specific challenges that the groups are facing, with a lot of hands-on practice with examples that are relevant to their work.
Another format this takes is aimed at individuals in public workshops, where anybody can sign up and spend the day learning how to do this both through a little bit of theory and a lot of practical strategies and hands-on examples. I do those in various cities mainly throughout the U.S., but every once in awhile in Europe and other places as well.
I’ve written a book on the topic called Storytelling With Data, which brings together a lot of the lessons that I cover in the workshops and goes deeper with more examples, tips, and strategies. I also aim to teach through my blog at StorytellingWithData.com.
What would you say shaped the perspectives that you currently hold around data visualization and what set you down this path?
My career started out in banking in credit risk management before the subprime crisis and thus before anybody really knew what credit risk management was. Banking was a fantastic industry to start out in—it was the first place where I really started to see the value of data done well when it comes to helping people make smarter decisions or understand something better.
From banking, I went into a space that’s much less historically known for being data-driven, human resources, but I joined the human resources team at a very data-driven company, Google. I joined the People Analytics team at Google, which is an analytical team embedded in Google’s HR organization. I was using a lot of the same principles and strategies that I had used to understand data and banking, but applied to a totally different topic—people—addressing things like, how do we engage employees? How do we understand how effective our managers are? How do we predict when somebody might be going to leave the organization?
Also, while at Google, I had the opportunity to build out a course on data visualization, an area I’d been really interested in. It’s like I said before, when you make data visual, you make it accessible to more people. This gave me the opportunity to pause, do some research, and understand why some things that I’d arrived at through trial and error over time were effective. And, in terms of specific influences on the course, certainly Edward Tufte and his books on data visualization and Stephen Few, especially Show Me The Numbers helped.
So, I built out that course, taught it at Google offices throughout the world, and then started getting inquiries from other organizations saying, “Hey, we’ve heard that you do this. Can you come do it for us as well?” I’ve since been spending my time in the past four years doing related workshops, the book, and the blog.
In your book, one of the first things that you mentioned was that generally we’re not very good at storytelling with data. For all the buzz in 2016 around using data in business, why do you think we’re still fundamentally lacking in this way?
Nobody really teaches us how to do this and yet, as we amass more and more data, there’s a growing desire to make sense out of all of it. Visualizing it is one way of turning data into information, then telling stories with that information helps people understand and act upon it. Historically, these aren’t skills that have been taught.
There’s a growing desire to make sense out of all of this data. Visualizing it is one way of turning data into information, then telling stories with that information helps people understand and act upon it.
You think back to school, and you spent a lot of time learning about language and about math, but the two were never paired. Right? Nobody really teaches us how to bridge the two. There’s a ton of strategic advantage to be gained by individuals and companies who can do this well.
How would you describe the difference between data visualization and what you would call storytelling with data? Because I feel like a lot of people will look at a chart and they’ll say when they’re generating a report, “Oh, I can generate a graph. It can’t be that complicated,” and then they copy and paste it and present it. Where are we going wrong with all of that?
Data visualization takes a lot of different forms. We often use data visualization as we’re exploring data to try to figure out what might be interesting, what can we learn from it, what might somebody else care about.
For me, data storytelling comes in when we’ve figured out those interesting things that somebody else may care about, and now want to explain it and really hone in on what actions we should take based on this and what we can learn from it. It’s easy to show data, and for an audience, it’s then easy for them to say, “Oh, that’s interesting,” and then move on to the next thing. If you’re telling a story with the data, instead, you’re engaging them. You’re asking for a specific action. Then, the audience has to respond to that.
It’s easy to show data. If you’re telling a story with the data, you’re engaging them. You’re asking for a specific action.
Even if they disagree with the story you’re presenting, it starts a conversation that may not ever happen if we’re simply showing data.
Could you explain the difference between exploratory and explanatory analysis?
That’s how I tend to think about some of the different parts of the analytical process. Exploratory analysis is where you start with a hypothesis or a question, and are exploring data, looking at it through different lenses, trying to understand, what can I learn from this data that somebody else might care about?
Once you’ve identified what somebody else cares about, then we move into the explanatory space, where you have something specific you want to communicate to somebody specific. This is where the idea of telling a story with data comes into play.
When you look at a data visualization, what would you say makes it effective or ineffective? Would you say there is such a thing as a “good” or “bad” method for visualizing data?
When I’m looking at a data visualization that’s meant to play a role in this explanatory part of the analysis that we talked about, I’m looking for a few things. I’m noticing where my eyes go first—do they go to the most important parts of the graph or the message? I look for affordances in design, so things like, is everything well-titled? Do I know what I’m looking at or do I have questions?
For me, the best data visualizations have a story to tell and give the audience a clear understanding of what’s important, what they should pay attention to, and why.
Two versions of the same graph, original (top) and a focused data visualization redesigned by Cole (bottom), via Cole’s post, Improving Upon “Good Enough”
If I needed to sit down and complete a report or create a presentation showing the impact my work has had, could you walk me through where I should start?
I’ll break it down to five core steps. These are the five core lessons that I cover in my workshop and that are also covered in-depth in the book.
- Have a robust understanding of the context. Before you spend a lot of time visualizing data or creating content, who is your audience and what do you need for them to know or do?
- With your data, think about what you want your audience to do, and choose a visual. Is there something interesting about the shape of the data? Do you want for them to compare two things? The answers to these questions will help you identify which visual may make it really easy for your audience to understand your message.
- Look at the visual you’ve created and identify clutter. Is there anything that’s there that isn’t adding information, that doesn’t need to be there? Strip those unnecessary elements away. Think about where you want to draw your audience’s attention, where you want for them to focus, what the most important parts of what you’re showing are, and make that obvious.
- Be strategic with design. Color, the size of different elements, and words can all help create a visual hierarchy, or implicit cues for your audience to help them know the order in which to process the information that you’re giving them.
- Tell the story. What’s the narrative that goes around this? How do I not simply show data, but rather make data a pivotal point in an overarching story?
What’s the narrative that goes around this? How do I not simply show data, but rather make data a pivotal point in an overarching story?
For me, once you’ve gone through those steps, you’ve considered the context, chosen an appropriate display, identified and illuminated clutter, drawn attention where you want your audience to pay it, and told a story. You’ve turned a graph into information that’s going to be relevant and useful for your audience to really drive them to action.
You talk about the “slideument,” which is this combination of document and slide deck. Do you have any tips for after you go through this whole process, when you’re making the final product and have to present it both as a presentation, and also send it out after the fact? How do we avoid the slideument?
This is a common challenge, where you’ve done an analysis, and now, you’re creating the presentation or the documentation of it, and want to be able to use the thing that you create—which is often, but not always a slide deck—to both present live as well as serve as the report-like thing that you send around.
As you can imagine, it’s hard to meet both of these criteria simultaneously because ideally, you’d have two different work products here. Your slides that you’re going to present would be very sparse because you’re there to narrate and to react to your audience and go deeper or less deep or speed up or slow down or answer questions as they come up, versus the report version that’s sent around where you’re not there to defend it or answer questions, so it needs to be more dense.
Oftentimes, this work product ends up being one and the same because of time constraints, so we just put together one deck that’s meant to serve both of these purposes. Inevitably, the content ends up being way too dense to put up on the big screen and sometimes not dense enough for the version that gets sent around.
There are a couple of strategies that you can use to try to overcome this. First, use animation in your live presentation. Instead of starting with the really dense slide or the really dense graph, I build it up piece by piece, enabling me to have different elements, different pieces of the data up here so that I can talk through them and have my audience looking exactly where I want them to look as I’m doing that. By doing this, by the time you build up to the final graphic, it no longer feels as complicated or as dense as if you were to just put that up all at once. There are some good examples both in my book and on the blog of this.
If you’re in PowerPoint-land or using some sort of presentation software, there’s the summary notes pane that you’ll see underneath the slide, and so you can think about keeping your slides sparse, but then putting the voice over, the narrative that you’ll be talking through in the live presentation. Put that down in that summary notes pane so that for the version that gets sent around, all of that context is still there to help people understand the story and understand what they should be paying attention to and why.
There’s another really common form of data visualization, the infographic. From a data storytelling perspective, and I know that maybe designers and journalists might have a completely different opinion, when would you say they’re appropriate and is there anything different in how we should approach them?
Infographics really run the spectrum. There are a lot of great examples in data journalism of really dense, interesting, inviting infographics. The New York Times does a great job with these. On the other end of that, there are the fluffy posters, where there’s a lot of data or facts on a given topic, seldom cohesive, and sometimes, when you peel back the data, there actually isn’t even much data underlying them.
For me, with infographics as with any sort of data visualization, it comes back to thinking about, who is your audience and what are you trying to do? What do they need to do? What do they need to know, and is an infographic going to help you get that across?
The best infographics, just like the best data visualizations, tell a story. Keep the audience in mind and walk them through a story so that by the time they’re done taking in the infographic, they have a better understanding of the topic.
There are tons of tools to visualize data. Do you have any favorites that you usually recommend for people who are maybe first-timers?
That’s actually a really common question I get. There’s no magic bullet, no tool is going to work perfectly. Any tool can be used well and any tool can be used not so well, so my view is always pick a tool, get to know it as best you can so it doesn’t become a limiting factor when it comes to doing some of the things that we talk about for directing your audience’s attention and visualizing data in a way that’s going to be easy for them.
Subscribe to The Analyst
Have feedback or suggestions for future episodes? Let us know here.