making a case for stacked bars

When I review common types of visuals used in a business setting during my workshops, one of the graphs we discuss is the stacked bar chart. I typically say something like the following:

“While we’re on the topic of bars, let’s talk about another common bar chart: the stacked bar graph. Stacked bars work well when you want to compare totals across different categories, and then within a given category, you want some understanding of the subcomponent pieces. Notice though, that they work less well if you want to compare those subcomponent pieces across categories. This is because as soon as you get past the first series, you no longer have a consistent baseline to use to compare. This is a harder comparison for our eyes to make, so something to keep in mind when reaching for stacked bars.”

I’ve noticed lately, though, that when a client shares a stacked bar as a makeover candidate, it usually gets remade into something else. Which has me pondering… Is there really a good use case for the (non-100%) stacked bar?

The most common scenario in which I see stacked bars used is to show total over time and also the change in composition. For example, I was just looking at a workshop example like this that depicted revenue over time: the overall height of the bars represented total revenue and then each bar was subdivided into source of revenue, in an attempt to show how the source of revenue was shifting over time as overall revenue changed. Here’s a quick sketch of what it looked like:

Stacked bar sketch - before.png
 

To me, this was a clear case where, by attempting to answer too many questions with a single visual, we don’t answer any one of them as well as we could if we broke it up into multiple visuals. In this case, my recommended view had a line graph showing total revenue over time, and a 100% stacked bar chart to show the relative shift in composition (revenue source). This is roughly what the visuals looked like (the real version was paired with additional explanatory text and emphasis that I've omitted for simplicity):

Stacked bar sketch - after.png

There is a challenge that frequently arises with the stacked bar: if anything interesting is happening further up the stack, it becomes challenging to see it because it’s stacked on top of other things that are also changing. This means potentially important components of the data or what we can learn from it can get lost or missed. In the above scenario, for example, there was an interesting shift happening in source of revenue over time that was hard to see in the original graph.

I’ve been racking my brain for good examples of stacked bars to figure out whether I should change how I discuss their use. I’ll ask for your help on this front momentarily.

I do have a couple of examples that are top of mind. There is a horizontal stacked bar that I highlight in Chapter 6 of my book as a model visualization:

In the above example, the most important thing is the length of the overall bars. It’s interesting to know what the subcomponent pieces are as a proportion of the given bar, but there isn’t a strong need to be able to compare those subcomponent pieces across bars. I think this works. Though I should mention that I’ve also received feedback (a one-off comment, so no idea how representative it is) that this particular visual is confusing.

As I write this, it occurs to me that I did also use a stacked bar in my prior blog post. This was a case where I wanted the audience to focus on the stacked piece (there were only two data series stacked on each other in this example), but the big picture opportunity the stacked portion illustrated across the various bars was more important than specific comparisons between the bars.

With that prelude, I'll turn the conversation over to you—have you seen examples of stacked bars that are effective? These can be vertical or horizontal (I think it's coincidence that the two I highlight above that work are horizontal and the one that didn't as well is vertical, but perhaps that's not the case?). Also, I limit my question here to the non-100% versions, as I do think there are more common use cases for 100% stacked bars, since you get additional flexibility with multiple baselines to align by and compare across (top and bottom-most data series in vertical 100% stacked, or left and right-most in horizontal 100% stacked). But I’m struggling to come up with many great use cases for simple stacked bars.

Please share your thoughts and examples by emailing them to stackedbars@storytellingwithdata.com by Wednesday, 11/22. Is there an example from your work where you’ve used a stacked bar effectively that you can share? (Please don’t share anything confidential—anonymize as needed.) Have you seen good examples in the media or elsewhere? Are there use cases you can imagine where a stacked bar would work well? What considerations should we keep in mind when using stacked bars? It will work best if you can share a visual, even if it's a simple sketch like I’ve included above.

I will pull together what I receive into a follow-up post. Stay tuned on that front and in the meantime I look forward to hearing from you RE: stacked bars!

is there a single right answer?

I had the pleasure of lunching with Stephen Few last week. His books (particularly Show Me the Numbers) and overall approach to visualizing data has influenced my work in a big way over the years. We had an enjoyable conversation (and some amazing caramelized brussels sprouts!) at a lively Berkeley cafe.

We agree on a lot of things. We spent some time bonding over a shared distaste for pie graphs. In our work, we both promote clutter-free graphs that focus on the data, with sparing and intentional use of color to direct attention. We also happen to both be introverts who are passionate about data visualization in a way that leads us to break out of our respective comfort zones and share our learnings with others.

But there are certainly also areas of disagreement. Which is where the real fun comes in. As we were preparing to leave, Steve asked if I'd be open to discussing some of these areas. Of course! Friendly debate can help to refine one's point of view, articulate their logic, and perhaps even change their way of thinking. 

One area of disagreement is on the use of 100% stacked bar charts. Steve already wrote a great post on this that summarizes our conversation well, which you can view here (be sure to read the comments, where our conversation continues and others chime in with their viewpoints).

Through some of our back and forth on 100% stacked bars, another area of varying beliefs came up. It has to do with the existence of a single "right" or "best" answer when it comes to visualizing data, the topic of this post. Let me summarize our respective points of view and then I'll ask you to share yours.

I believe that data visualization sits at the intersection between science and art. There is absolutely some science to it: best practices and guidelines to follow. But there is also an artistic component. This means that two different people may approach the same data visualization challenge differently. There is room for diversity of thought and approaches. But it is important that we use our artistic license to make the information easier for our audience to understand, not more difficult.

My view is that we are often faced with tradeoffs when it comes to determining how we design our visuals. For example, I may choose a 100% stacked bar even though it makes it hard to directly compare two segments within the bar (assuming that is a lower-priority comparison and it allows me to still make my main point clearly) because with it, I get an implicit visual cue that the segments are part of a whole. I've decided that the tradeoff is worth it. Or I may include an annotation on my graph, even though it adds what someone else might consider clutter, because I believe it makes an important point in the data easier to interpret. I might choose sparing use of bright orange to draw attention to a negative aspect of the data because it comes from the given brand's color palette, while someone else might choose black or red for the same purpose. In these situations, none of these choices is necessarily the "right" choice. Different people will make different decisions. I think that is ok. For me, paramount is that design choices like this are intentional—I considered the tradeoffs and made a decision in light of those—not happenstance. Yes, they should be informed by science. But my view is there is also room for personal choice.

Steve's view is that the introduction of art into this conversation is a slippery slope because individuals will define art differently. He states, "there is no art in what I do; it's all science" and goes on to say that when he makes a judgement to design a data visualization in a particular way, he's basing that judgement on a fairly deep understanding of what works and what doesn't, which is rooted in science. That said, he concedes that there is a creative component but that this creativity and judgement should always be applied in an effort to create the most effective outcomes in understanding.

Given this, his view is that in a particular situation, there is a single right or best answer for visualizing data. In a comment on his blog, he poses the following question to me:

In your opinion, are some data visualization solutions better than others in a given situation or are they all equal in merit? If the former, on what basis can we judge one to be better than another? You mentioned that data visualization designers should “make the information easier for their audience to get at, not more difficult.”? This suggests that we can judge the merits of a data visualization by its ability to make the information as easy to understand as possible. Choices of graph type, colors, etc., are made to present the information as clearly and accurately as possible. If one data visualization does this better than another, it is the better solution–correct? If this is true, does it make sense to say that there is no “best” solution in a given situation among a set of proposed solutions?

I don't disagree that some data visualizations are better than others. We've certainly all seen examples of data visualizations that are inferior for a great number of reasons (bad chart choice, lack of clear labeling, too many colors, unnecessary clutter), while others are more effective. For me, though, it is possible to have multiple varying visuals that may be equally effective (or roughly equally, if the differences are marginal enough not to matter and all do a good job of getting the intended information across).

What is your view? Is there a single "right" or "best" answer when it comes to data visualization, or is there room for varying approaches? Leave a comment with your thoughts. I look forward to the continued debate!

500,351

That's the number that just caught my eye when I was looking up a past post on my blog a moment ago. If you look over to the left, you'll see it, too. Well, not exactly, as it's likely continued to tick up a little since I began writing this post. It reflects the number of views this blog has had since I launched it in December 2010. Just two words come to mind when I see it: thank you.

Thank you for reading, for your comments and your general interest in storytelling with data.

This seems like a good (albeit random) point to pause for a moment, to link back to some popular posts you may have missed and ask you to weigh in on what you'd like to see here in the future.

popular posts in case you missed them
The 10 most popular posts (based on number of page views) are listed below (plus a bonus 11th that has 10 tips and links to related posts).

  1. no more excuses for bad simple charts: here's a template
  2. how to do it in Excel
  3. the waterfall chart
  4. strategies for avoiding the spaghetti graph
  5. a Google example: preattentive attributes
  6. my penchant for horizontal bar graphs
  7. chart chooser
  8. the power of simple text
  9. logic in order
  10. slopegraph template
  11. celebrating (almost) 100 posts with 10 tips

what would you like to see covered here in the future?
Suggestions on future topics, questions you'd like me to opine on, or data visualization challenges you're facing are welcome. Leave a comment with your thoughts or email me directly at cole.nussbaumer@gmail.com.

where should I go in 2015?
I'm also looking ahead to my 2015 public workshop schedule. If you'd like to recommend a city or location, you can do so here.

Thank you very much for reading!

a storytelling with data ad

One of my favorite indulgences on a weekend morning is to sit in the sun on our terrace and read the latest copy of Dwell magazine. A number of things in the universe have to align to make this possible: namely, the sun must be shining and the child must be sleeping. The universe aligned in just this way this past Saturday (bliss!).

I find that the design of products and the design of spaces can sometimes influence my thinking, spark an idea, or act as inspiration when it comes to the visual design work on which I focus much of my attention. On this particular read through Dwell, it was the following advertisement that caught my eye:

This ad caused me pause for a few reasons:

  1. The leading stat - 1 in 5 children go to school hungry - is powerful. When it comes to communicating a number or two, tables and graphs don't usually have a place, as the numbers themselves carry a lot more attention-grabbing power.
  2. The use of preattentive attributes to make certain elements of the visual distinct: the numbers at the top are in bold, all caps and underlining draw attention to the second line, and the sort of sea-green in the logo and text at the bottom emphasize the un and is (when it comes to this last point, I might have chosen different portions of text to draw attention to, but I think that's one of those things that can be up for debate and probably there was a good reason the designer chose these particular pieces - perhaps the dichotomy between un and is?).
  3. The story. It's short and sweet, but still a robust example of storytelling with data, which, with the personal anecdote and picture are made to be much more human than a simple stat on its own would be.
  4. The picture. Speaking of pictures, one frequent question in my workshops is about the use of pictures when it comes to visual communication. I don't use pictures a lot personally, but as mentioned above, I do definitely think there are ways to use pictures that appeal on a different level than numbers do. Here, I think the pairing of the two is effective.

What do you think of this ad as an example of storytelling with data? Is it effective? Why or why not? Leave a comment with your thoughts!

slidedocs

In my workshops, we often discuss the challenge that arises when the communication you're creating is meant to be both 1) projected or handed out in a live presentation, as well as 2) sent as a follow up for those who attended to refer back to, or to fill in those who weren't able to attend the presentation. In an ideal world, these would be two separate work products. In reality, however, this often gives rise to what Nancy Duarte calls the slideument, a single document that is meant to address both needs.

The problem, of course, is that by trying to meet too many needs at once, the slideument doesn't address any of them perfectly. If you're interested in my further pontification on the topic and some thoughts on how to address the challenges that arise with the slideument, you check out this blog post.

My interest was piqued when I saw that Nancy Duarte recently introduced a new concept for using slides, which she has termed Slidedocs. The description on her site describes slidedocs as "a visual document, developed in presentation software, that is intended to be read and referenced instead of projected." The FastCo Design article written about Duarte's slidedocs is titled "Book Written Entirely in Power Point Aims to Reinvent How Businesses Communicate" (article).

With my busy schedule this week, I haven't had a chance to give the 150 page doc more than a cursory review. But rather than wait until I have more time to consume and reflect before sharing, I figured I'd post this now and gather your inputs. What do you think? Will this revolutionize the way we communicate? Leave a comment with your thoughts. I'll add mine once I have a chance to look more closely at the detail.

failure in design(er)

Yesterday evening, our recently purchased, lovely new couch arrived. Or, rather, the large boxes that contained our recently purchased, lovely new couch arrived. Suddenly, it was very clear what we gave up by not springing for "white glove delivery".

Not to fear, though. It came with instructions. My husband and I can both read and follow instructions.

Right?

Easier said than done, it turns out. These were certainly not the worst assembly instructions that I've ever seen, but they left a lot to be desired. Perhaps a very lucky or clever individual could get it right the first time (we were neither of those, as it turns out). But you'd have to know which details were important to pay attention to.

We had several false starts, turning the diagram round and round to say: Ah, now I get it! Wait, no, now the one frame piece is too long. Oh, now I see the problem. Oops, no, now the holes don't line up. After several such instances, we recognized that the bars in the frame are not equidistant apart (and it matters which two are closest together), we realized that two of the frame bars had four holes each and the third had two holes each and that the relative positioning of the bars with respect to one another is important, we learned that FX1 and GX1 are in fact not interchangeable (even though at the top they're shown with FX1 clearly on the left and GX1 on the right, but then below are less prominently switched).

Now that we've assembled the couch correctly (finally), we could do it again without breaking a sweat. We know exactly which are the important parts in the diagram to pay our attention. But why was it so difficult the first time around?

I'm in the middle of a book I'm enjoying, The Design of Everyday Things. In it, Donald Norman asserts that when you have trouble with things, you shouldn't blame yourself (even though that tends to be people's natural tendency). Rather, it's the fault of the design and you should blame the designer. While this book focuses mainly on product design, I think many of the insights are true in the data visualization space as well. In this case, the corollary is clear: if you are struggling to understand a visual representation of data, don't blame yourself; blame the designer. Odds are, they didn't adequately take your needs as the audience into account in their design process. For those designing visual displays of information, this is a reminder to always keep your audience in mind, for, as Donal Norman says, "well-designed objects are easy to interpret and understand."

I unabashedly blame the designer of the instruction diagram for our difficulty assembling something that could have easily been straightforward. If the designer had thought about the intended users and leveraged affordances to make it clear which details were important and should be paid attention to, my husband and I would have had a much less frustrating process assembling our (now truly lovely) couch.

What design issues cause you frustration? What can we learn from this to apply in the world of data viz?