the cat in the hat knows a lot about data visualization

I recently had the pleasure of guest lecturing a Stanford MBA class, Strategic Communications. Here, I've recorded a 20-minute segment from that lecture, which covers two basic things that you should do when communicating with data:

  1. Be sparing and intentional in your use of color, and
  2. Put your thoughts into words.

Check out the video below for some quick lessons and examples. Thanks, JD,* for inviting me to share with your class!

Check out my YouTube channel for more videos.

*JD Schramm lectures in the Knight Management Center in the Graduate School of Business at Stanford University; check out his recent presentation, The Secret to Successful Storytelling with Statistics.

/
CLICK HERE TO JOIN OUR MAILING LIST

SEARCH STORYTELLING WITH DATA: © 2010-2016 Cole Nussbaumer Knaflic. All rights reserved. STORYTELLING WITH DATA and the STORYTELLING WITH DATA logo are trademarks of Cole Nussbaumer Knaflic.

hello baby

It has been quiet here for a bit and after this post will likely continue to be for a while. Why?

Two words: baby Eloise.

Like her oldest brother, Avery, Eloise surprised us early, spending several weeks in the Neonatal Intensive Care Unit (NICU) before finally coming home. 

Eloise was born across town at the same hospital as Dorian. Hospitals, as it turns out, are full of data visualizations. We were at the hospital for some routine tests when the adventure began. The machine they hooked me up to drew data on paper. While to me, it just looked like a lot of peaks and valleys, apparently to the doctor, it depicted "active labor." Interesting. But we only looked at that particular graph for a matter of minutes. It was the following ones that we stared at for weeks on end in the NICU, alarms periodically blaring.

In the first image, the bottom graph (right hand side) shows contractions every minute. The second image is a snapshot of ventilator stats. The final image depicts the stats consistently tracked while in the NICU: heart rate (green), oxygen saturation (blue) and respiratory rate (white).

In the first image, the bottom graph (right hand side) shows contractions every minute. The second image is a snapshot of ventilator stats. The final image depicts the stats consistently tracked while in the NICU: heart rate (green), oxygen saturation (blue) and respiratory rate (white).

Spending time in the NICU is a strange experience. The ups and downs are of course scary (when I learned that the treatment for the first several days wasn't working and Eloise would have to be intubated and put on a ventilator was one of the worst moments of my life). There's also a strange tension of emotions. On the one hand, you have professionals taking amazing care of your baby around the clock. But it's difficult every single time you have to leave and aren't able to take your baby with you. Being there is stressful. There's a feeling of guilt, though, whenever you aren't there. Having been through the process once before made it familiar, but not any easier. The day we were finally able to take Eloise home was a glorious one.

And now that Eloise is here with us, the data visualizations don't end. There is the temporal data I'm recording in a list (old-school-style, with my fancy tools of a pen and spiral notebook) on feedings and diapers. There's the Jawbone UP app on my phone, a daily reminder of my interrupted sleep and how little sleep and steps I'm getting in general. There's the automatic graphing that our high-tech scale does of my weight (just what every recently pregnant person wants to see, right?!?). Even one of the bottle packages had a graph on it! 

The first two images are from Jawbone UP, which tracks my sleep and steps; the first image depicts a night of sleep (and wakefulness, shown in orange—feeding times) and the second image shows my total sleep and steps for a given day. The third image is a screenshot of my weight over time collected via our Withings scale (y-axis scale/labels intentionally not shown; if I could annotate the peaks, the first would read "Dorian birth" and the second "Eloise birth," perhaps I'd also draw a "goal" line somewhere near the bottom!). The final image is from a package of Dr. Brown bottles—great use of color and text in graph to highlight the Dr. Brown line.

The first two images are from Jawbone UP, which tracks my sleep and steps; the first image depicts a night of sleep (and wakefulness, shown in orange—feeding times) and the second image shows my total sleep and steps for a given day. The third image is a screenshot of my weight over time collected via our Withings scale (y-axis scale/labels intentionally not shown; if I could annotate the peaks, the first would read "Dorian birth" and the second "Eloise birth," perhaps I'd also draw a "goal" line somewhere near the bottom!). The final image is from a package of Dr. Brown bottles—great use of color and text in graph to highlight the Dr. Brown line.

There are certainly many more stats that I could be tracking and visualizing. But I'm not going to. Rather, I'm going to spend my time staring at this beautiful, tiny creature.

Her loving big brothers and father have been doing the same.

Welcome, Eloise, we are so very happy you are here!

Eloise Noel Knaflic
Born February 19, 2016
5 pounds 3 ounces

/
CLICK HERE TO JOIN OUR MAILING LIST

SEARCH STORYTELLING WITH DATA: © 2010-2016 Cole Nussbaumer Knaflic. All rights reserved. STORYTELLING WITH DATA and the STORYTELLING WITH DATA logo are trademarks of Cole Nussbaumer Knaflic.

declutter your data visualizations

When I was a little girl, I used to get in trouble for cleaning my room. Check out the following video to find out why and learn why clutter in data visualization is a bad thing and how to avoid it. Specifically, I'll cover five tips and examples from my book, storytelling with data

  1. Leverage how people see
  2. Employ visual order
  3. Create clear contrast
  4. Don't over-complicate
  5. Strip down & build up

This is a slightly modified version of the talk I've been giving on my Bay Area book tour at companies like LinkedIn, Facebook, Pinterest, Dropbox, Tesla, Airbnb, and Evernote. Post any related questions in the comment section and I will respond. I hope you enjoy!

/
CLICK HERE TO JOIN OUR MAILING LIST

SEARCH STORYTELLING WITH DATA: © 2010-2016 Cole Nussbaumer Knaflic. All rights reserved. STORYTELLING WITH DATA and the STORYTELLING WITH DATA logo are trademarks of Cole Nussbaumer Knaflic.

the use case for less common graphs

In my book, I write about the 12 types of visuals I find myself using most frequently for communicating data in a business setting (here is a list of the 12 with links to examples; here is a related post on storyboarding for the chapter on visual displays in my book and the final graphic depicting the various visuals). For the most part, the graphs I find myself turning to regularly aren't anything crazy. They are line graphs and bar charts—graph types with which people are familiar. There is a good reason for that: when you use a graph that people already know how to read, you face less of a learning curve with your audience for getting the information across. In other words, instead of spending their brain power figuring out how to read the graph, they can focus on what interesting or useful information they can get out of it.

But does that mean there aren't use cases for other types of graphs?

Certainly not.

My view is there is absolutely room to play and be creative when it comes to visualizing data. But novelty shouldn't be prioritized over effectiveness. In general, the goal should be to visualize the data in a way that's going to get your audience's attention and keep it long enough to get your point across.

To talk about this more concretely, let's look at some specific examples of less-familiar graph types. I'll put these into two categories: effective examples (I'll highlight a few) and ineffective ones (a single poor example from my own work). For each, I'll talk about some related general considerations when determining how to visualize data.


Effective examples of less common visuals

You should expect that with less common visuals, it will take the audience some time to understand what they are looking at. But in cases where this is done well, it shouldn't take very much time, because it's intuitive and/or due to the visual cues and text (or live narrative) you've included to ease this process of understanding. For me, a less common visual is successful if it doesn't take me long to "get it" and once I do, I'm left thinking "that is clever." 

Here are a few examples that meet the above criteria for me (click on any of the images for more details):

pitchinteractive.com

pitchinteractive.com

flowingdata.com

flowingdata.com

krisztinaszucs.com

krisztinaszucs.com


Ineffective example of less common visual

For an example failure when it comes to using a less common graph type, I'll look to my own work. In a prior life, I worked in banking managing home equity fraud. Fraud management is often discussed and measured in terms of the eight stages of the Fraud Management Lifecycle: deterrence, prevention, detection, mitigation, analysis, policy, investigation, and prosecution.

To show progress in each of these areas for a number of different types of loan fraud, I though the spider (or "radar") graph would be perfect: the greater coverage on the graph, the greater coverage we had in a given area. This would help us understand and highlight strengths and weaknesses, which could be used to help determine priorities going forward. I created a visual similar to the following:

I wanted this visual to work very badly. It seemed to me that it was the perfect use case for the spider graph. But I found I was having to spend a significant amount of time explaining how to read the graph each time I used it. Even after doing so, I wasn't having great success getting my point across. I ended up creating a graph to explain how to read the graph (note that the last effective example shown above does this as well, but in a quick-to-understand way, so I think it actually works in that case, though it didn't in mine).

At this juncture, I finally realized that my visual was failing. It wasn't intuitive to my audience—no matter how much I wanted it to be!

Trying to force my approach on my audience was not a recipe for success. Rather, it actually distracted from the message I wanted to communicate. So I had to go back to the drawing board and figure out what I most wanted my audience to be able to do with the graph, then brainstorm and iterate to find a way to get that information across more intuitively. Here is a related post with an alternate view of the data using one of my go-tos: the horizontal bar chart.

When you find yourself fighting against your audience in the way you're visualizing the data, it isn't working. Don't blame your audience. Blame your design.


Meta-lesson: audience is king! Don't be afraid to try new or novel ways to visualize information, but be willing to reassess and change approach if needed. Your main goal when communicating with data is to get your message across to your audience, so chose an approach that will allow you to effectively do this!

If you are aware of other effective examples of less-common visuals—there are certainly many more out there—please leave a comment with the link.

/
CLICK HERE TO JOIN OUR MAILING LIST

SEARCH STORYTELLING WITH DATA: © 2010-2016 Cole Nussbaumer Knaflic. All rights reserved. STORYTELLING WITH DATA and the STORYTELLING WITH DATA logo are trademarks of Cole Nussbaumer Knaflic.

strategies for avoiding the slideument

A popular question posed to me (in general and several times just this week) is: what should I do when my deck is meant both to meet my presentation or live meeting needs and will also be sent ahead as a pre-read or afterwards as a follow-up?

This scenario often leads to a slideument—part presentation and part document, and not exactly meeting either of the above needs. The slideument is typically too dense to put up on the big screen and often not detailed enough for when you aren't there to talk through the material. I've written about this challenge before here.

Here, I'll suggest three potential alternatives for avoiding the slideument:

1. Create two different documents. Ideally, these two situations call for two totally separate work products: sparse slides for when you are talking through the material live and a denser, more detailed report-like document for the version that is disseminated. If it's something really important, consider whether creating two separate documents makes sense.

2. Leverage animation & annotation. If you're working in PowerPoint land (or something similar), you can animate a sequential appearance of elements on a slide, focusing your audience's attention exactly where you want it when you discuss the material live. Then the sent around version would have the final fully-built slide or be a version that annotates via text what you would say in the meeting or presentation. This blog post illustrates this approach using a specific example (from my book).

3. Make use of the Notes feature. If you're working in PowerPoint, make use of the pane below each slide that says "Click to add notes." Leave your slides sparse and put the narrative that you would say for each, or any additional context that is needed, in words in that notes pane. Just make sure to alert your audience of the sent-around-version to look there for details.

Do you have other ideas on how to address this challenge? If so, please leave a comment with your thoughts or suggestions on what you've seen work well.

/
CLICK HERE TO JOIN OUR MAILING LIST

SEARCH STORYTELLING WITH DATA: © 2010-2016 Cole Nussbaumer Knaflic. All rights reserved. STORYTELLING WITH DATA and the STORYTELLING WITH DATA logo are trademarks of Cole Nussbaumer Knaflic.

be gone, dual y-axis!

Due to popular demand (and a growing waitlist for the sold out workshop on 2/3!), I've scheduled another upcoming workshop in San Francisco on 2/8. Details & registration here.

I am generally not a fan of the dual y-axis. I've written about it before and offered some alternatives in my book and here. When scrolling through my Twitter feed last night, I came across the following offender and couldn't resist trying to improve upon it and taking part in the #MakeoverMonday challenge.

Before we get to the graph, let me spend a moment on #MakeoverMonday. This is a weekly segment that Andy Kriebel has been doing for quite some time (he's a busy guy; check out another fun current project of his with Jeff Shaffer, Dear Data Two). The premise is simple: take a less than stellar graph and make it better. In 2016, he's adding to the fun by doubling the Andys (having Andy Cotgreave join him) and also opening it up to the public. I think this is an awesome way to get people involved, share best practices and ideas, and generally inspire.

OK, back to the graph. The main challenge with a secondary y-axis is that there's always some questioning up front about which data to read against which axis. This particular example isn't horrible in that regard—the left and right orientation of Online hotel revenue and Number of travel agents titles, respectively, make it fairly clear which axis is which (though it does feel a little strange that each title is closest in proximity to the other data series, not the one it describes) . This graph does have another issue introduced by the secondary y-axis, however: the appearance of a crossing of the lines between 2005 and 2006. This looks like it might be something noteworthy, but actually is only a function of the scale used on the axes that creates a condition that they happen to cross each other at that point. Different scales would have them crossing in different places. I'd argue that they shouldn't cross at all.

In my workshops and book, when the topic of the secondary y-axis arises, I generally focus on two alternatives: 1) not showing the second (right-hand) y-axis but rather labeling the data in the secondary series directly or 2) pulling the graphs apart vertically so you can still leverage the same x-axis across both, but each gets its own left-hand y-axis so you can title and label them directly. Today, I thought I'd focus on a third potential alternative: turning the data that would be on two separate y-axes into the same units so you can simply plot it all on the same axis.

"Thought" in the preceding sentence is key. I'd envisioned my solution and penned the majority of this post before graphing the data (I've been doing this long enough that I should have recognized the danger in this). I thought I had the perfect solution in mind, but then graphed it only to recognize, "oops, that doesn't work." So let me rework the rest of this post. I'll take you briefly through my failed iteration and thought process as I do so.

Back to the idea of making the units the same and plotting it all on a single axis: this won't always be possible or appropriate, but I think (thought) it may work well in this case. For me, the point of this graph is that online hotel bookings have increased hugely over the past 15 years and that this has been—understandably—accompanied by a marked decrease in the number of travel agents. Since we're talking about increases and decreases here, one way to tackle would be to transform the numbers into relative increases and decreases and plot those directly. Here's what that could look like:

I hadn't looked at the numbers closely before graphing this, so failed to realize that the increase in online hotel bookings waaaay outpaces the decrease in travel agents. This totally makes sense now that I pause to think about it. But before seeing it, I was imagining a graph where online hotel bookings would be going upward to the right (as they are) and travel agents would be following perhaps a similar trajectory but downward to the right. The issue is that when scaled properly, the percent decrease in travel agents is totally dwarfed by the increase in online hotel bookings, so you don't really get a lot of value from the slightly downward sloping line (which also gets covered up in an unideal way by the x-axis labels).

Before seeing this, I was planning to discuss how moving from real numbers (for example, revenue or number of agents) to a percent (in this case, % change) causes you to lose something (sense of scale of overall numbers). I was then planning to go on to show a couple different ways to overcome this—first, by adding numbers to the graph directly, second by showing the actual numbers over time as well in bars (with number of travel agents being plotted in the negative direction) but pushing the bars to the background so they add a bit of context without a lot of clutter and maintain focus on the percent change. But none of this discussion makes much sense now that my original graph doesn't work.

Rather, after seeing the numbers graphed and recognizing just how huge the increase in online hotel bookings has been over the past 15 years, I'd be apt to just focus on that. The travel agent decline can become more of an interesting tidbit, included through use of text (not graphed at all). It isn't exactly the eloquent solution I was imagining, but it's where I'm going to land this time:

Be sure to check out Andy K and Andy C's respective solutions (here and here) where they go through more iterations and potential solutions for reimagining the original dual-axis graph. Interestingly, they both ultimately landed on scatterplots. For me, this does something strange to the dimension of time, but with adequate labeling (or animation, as Andy C uses), perhaps this is overcome. Take a look and see what you think.

By the way, if you're looking to hone your data visualization skills, consider participating in a future #MakeoverMonday challenge. I'll be following along. I hope to see your contribution there!

/
CLICK HERE TO JOIN OUR MAILING LIST

SEARCH STORYTELLING WITH DATA: © 2010-2016 Cole Nussbaumer Knaflic. All rights reserved. STORYTELLING WITH DATA and the STORYTELLING WITH DATA logo are trademarks of Cole Nussbaumer Knaflic.