the slideument

A common question has come up in several of my recent workshops: what should I do when the document I'm creating is meant to be used both as a written report and as part of a live presentation?

In an idea world, this situation would never arise. Rather, you would prepare two distinct deliverables:

1. A written report, where you can get away with denser content and rely more heavily on things like written words and an appendix to make sure the necessary context and explanation are present for your audience.
2. A presentation, where slides are much less dense, font is never smaller than 16-point, and the speaker is able to verbally provide the necessary information so it need not all be physically written down.

In reality, this rarely happens. Time and other constraints lead us to create something that is meant to be a sort of mesh of these two things: the slideument.* 

*I can't take credit for this mashup, but rather must give it to Nancy Duarte, who discusses the slideument in her book, slide:ology.

So the question remains: if slideuments are the reality, what should we do? In this post, we'll take a look at some of the challenges this presents as well as some strategies for overcoming them.

The crux of the challenge is: the report needs to be able to stand on its own without a presenter there to explain it. But if you put the dense slide that meets this first need in front of your audience, you lose them immediately because they turn their attention to trying to understand what you've put in front of them and stop listening to you. Or worse yet, they see that what you've put in front of them looks overwhelming, so they tune both it and you out and turn their attention to something else altogether. In either case, you've lost some of your audience and thus your ability to communicate effectively.

One solution I'll propose is to maintain the density of information (to ensure an audience who is consuming this info on their own has the details necessary to do so) and use animation in your presentation software to enable to presenter to focus on just one piece of the visual at a time, while simultaneously ensuring that's where the audience's attention will be focused as well. In this way, we are able to lead our live audience through the pieces and communicate effectively to those digesting the information later with the written report.

Let's look at an example. Imagine that I want to show an overview of my organization's social media followers (assume this is context that sets up the interesting story that we aim to tell in the rest of the presentation/report). Imagine also that I've been given the constraint of a single slide to do this. My slide might look something like this:

While this level of detail might be fine to have on a single page of a report that someone looks at on their computer screen or on a printed page, it can be overwhelming when you put it up on the big screen. There's a lot to take in, so if you show it all at once, some in your audience may tune out entirely, while others will busy themselves reading through what they are looking at (and unfortunately, when they do this it means they've stopped listening to you).

One way we can prevent this is by only showing a single element of the page at a time. Tactically, you can do this either by having elements of the page appear and disappear or covering up elements on the page with white (either solid or semi-transparent) boxes and using animation in your presentation software to only show one item at a time. So perhaps you flash up the full slide as you tell your audience you are going to talk them through what they are looking at, piece by piece. First, focus on the membership trend over time:

When your audience can only see one element on the page, their attention will be on it and on you. After you've talked through the membership trend, you could shift to the next element: who the members are.

At this point, you could even layer on some preattentive attributes, like color, to draw your audience to a specific part of the element and talk about what's interesting there.

(Yes, the font is tiny; if you're there to explain why the blue part is interesting, perhaps the supporting text at the very bottom is best left as part of the report version and never presented). Then you could step through the remaining elements in the same manner, one by one, explaining and pausing to point out the interesting aspects of each.

Sure, it would be better if you could give each of these elements their own slide. But in the scenario where that isn't possible, perhaps you'll find these tricks helpful.

You can download the PowerPoint here. I've included a version where the graphs appear/disappear directly, as well as one that leverages transparent boxes to direct the audience's focus to a single element at a time.

data visualization: a reflection of personal taste?

When baby Avery took a slightly longer than expected nap earlier today, I found myself able to escape to one of my favorite spaces: the lounge chair on the patio in the sun with a copy of Dwell magazine.

I find inspiration in the design of physical things: spaces, furniture, gadgets. There was a paragraph in one of the articles I was reading that gave me pause -

"But there is a clear difference between the thoughtful selection of furnishings and the heedless accumulation of 'stuff.' Good furniture at its best is the intersection of art, craft, and industry. The 'right' pieces address the needs dictated by our behavior and predilections. That's the lofty ideal. Of course, complications arise when one starts calculating the pesky realities of space, time, and budget. And so begins the tradeoff." 

- Many Voices, Many Rooms. Dwell Magazine, Vol 13, 6/13.

I can see many parallels between what editor-in-chief Amanda Dameron says here about furniture and good data visualization. I guess it all comes down to thoughtful design. But personal taste must enter the equation as well, right?

Upon reading this, I started thinking about how my own personal taste when it comes to physical items is reflected in the data visualizations I create and in my stylistic approach to data visualization in general. In the spaces in which I live and work, my tendency is mostly minimalist. The color palette tends to be neutral, with sparing, bright splashes of color. No one who is familiar with my data viz style will be surprised that the dominant color is blue.

When I started thinking about it, I realized that my design of the physical spaces in which I spend time has many similarities with the data visualizations I create. So interesting!

I'm curious: how does your personal style influence your data visualization? Leave a comment with your thoughts!

Painting pictured above is by Jill Henry, ceramic bowl by Heath Ceramics.

plotting a value within a range

I often refer to my method of creating graphs as "brute force Excel," meaning that you can make almost anything work in Excel, but it sometimes means getting a little creative. The method described in this post can be considered a prime example of brute force Excel.

This example was taken from a recent workshop, where I've anonymized the details. The goal was to understand the cost of various high cost items in a given region relative to some comparison regions. The hypothesis was that the region of interest (Region 1, in the example below) had higher cost items than the other regions. The original visual was a table that looked something like this:

The table in this case might be helpful for the exploratory analysis, but it isn't well-suited for explanatory analysis, because there is simply too much to try to take in to form a conclusion. So I knew going in that I wanted to make the data more visual. The question was: how?

While I was considering this, I sketched out an idea:

Rather than show the individual values for each of the regions, my idea was to plot the range of prices across the various regions, from the minimum value to the maximum value (irrespective of which region has the min or max price, since it didn't seem like that was critical for what we are trying to show here) and then show the specific value for Region 1, the region of interest, within that range.

The next challenge was how to create this visual in Excel. I'll take you through how I did this. The brute force in this case is a couple of stacked bar charts, with some of the series made to be invisible. Here's the step-by-step:

STEP 1: Plot the minimum price and range (the difference between the maximum value and the minimum value across the various regions) as a stacked bar chart on the primary axis.

STEP 2: Make the blue series (Min) invisible by formatting it so there is no fill color, no line, and no shadow (I will never understand why Excel adds a shadow in the first place!).

STEP 3: Plot the Region 1 price and Marker (for this, I just put a value of 100 across all of the items - just big enough for it to show up on the graph - this will be what we'll use to show where Region 1 falls within the range) on the secondary axis. At this step, you must set the secondary axis maximum to $16,000 (overwriting the default of $10,000) so that it lines up with the primary x-axis.

STEP 4: Add data labels to the green series (Region 1) and format them so they are placed at "inside end" of the bar. Then make the green series invisible (no fill, no line, no shadow).

STEP 5: Format, format, format! The base visual looks like what we want. Now it's about playing with the details of the visual so that the information is as straightforward to consume as possible.

In this step, I think about getting rid of clutter, using color strategically to draw my audience's eye to where I want them to pay attention, and adding titles, footnotes, etc. to describe and explain my visual. A couple of notes on changes I made in this step:

• The item names and horizontal borders are in the cells directly (vs. part of the graph); I find it easier to format the visual this way, you just have to be sure that you've sized the graph so it lines up to the labels. 
• I reversed the order of the y-axis so the items range from min to max (vs. max to min as shown in the above steps) to leverage how our eyes tend to take in information (from left to right).

The above visual still takes a little processing to figure out what's going on. But I find the data much faster to process here than it was in the original table. Now, we can see relatively quickly that the price in Region 1 is mostly on the lower end of the ranges when compared to other regions (with a couple exceptions).

The final step would be to put the story around this visual: what it shows, why that's interesting, and what action should be taken (hard to do with the generalized example, but there was an interesting story here in the original version).

The Excel workbook with the progression outlined above can be downloaded here.