Make Your Data Pop « Thoughts On Presenting & Design

8Mar/094

Make Your Data Pop

Think what you will of him or his opinions, but Glenn Beck was the stereotypical presenter sharing data in tabular format on slides.

Some slides were more effective than others and some highlighted the perils of adding bullets after bullets or even rows after rows. You're forced to shrink the text and the message gets lost in the delivery.

Now, no matter what your data is, even if it has little economic, political, or scientific importance, it's critical to never manipulate the data. What you can and should do is format the data so that not only is it as truthful as it is in its rawest form, but that the underlying message is brought out more clearly.

With that said, let's look at a recreation of one of the slides he showed:

Within the wall of text, there's a message in there somewhere. OK, 13% believe that if you earn between $151,000 and $250,000, you're rich. Fine. Does that mean those same people think that if you earn outside that range, say $269,000, you're not rich? Of course not. That's a given, logically. Being "rich" is a state of positive wealth. You're not less rich if you're more wealthy. It's just impossible. It's a conclusion that couldn't be honestly debated even by the most semantically-obsessed individuals. Here's the key: If you're going to draw conclusions from the raw data, make sure you're on solid ground. If the data and conclusions were presented in a meeting or conference, provide the raw data in the handouts. Not only does it free the presentation from slides that make the audience work too hard to decipher, but it gives you a certain amount of transparency that shows you didn't monkey with the data to force it into the conclusion that you desired.

So, our first step is to get the data out of the table and into a chart. Since we're dealing with percentages, a pie chart is the way to go.

The pie chart using the data from the table really doesn't tell us anything that the table doesn't already tell us. It just gives the numbers a visual sense of scale. However, watching the video clip (starting at 4:41), Glenn makes his point that 20% of those polled thought those making $251,000 or more are rich. Well, based on the logic we observed above, that's not true. What we have here is not only a poorly designed slide, but an inability to read a data table.

If you apply the logic that anyone who thinks that someone making anywhere less than $251,000 is rich would also think that someone making $501,000 or more is rich, then you can create the following chart:

If you speak to Glenn's statement regarding the number of people who thought you're rich if you make more than $251,000, then you'd have the following chart:

That's a different percentage than he gave, isn't it? Well, as I said, he apparently has trouble reading data tables. What we have here as well is a slide that makes the point he was trying to make much more clearly and with more impact. Applying the same logic to the other poll answers yields:

If you want to make your data really pop and help you drive your message home, consider and focus on your message. Display the data in a way that reinforces your message, yet maintains the integrity and fidelity of the data.

Tagged as: charts, Slide Design, tables Leave a comment

kevinrussellersel

All employees are treated as if they are a crucial component of the company's success. Our owner and managers are very skilled at fidelity 401k making employees feel valued, respected, and appreciated. Despite long hours and challenging work, job satisfaction is very high because of this approach.
annktrembley

That’s a given, logically. Being “rich” is a state of positive wealth. You’re not less rich if you’re more wealthy. It’s just impossible. It’s a conclusion that couldn’t be honestly debated even by the most semantically-obsessed individuals. Here’s the key: If you’re going to draw conclusions from the raw data, make sure you’re on solid ground. If the fidelity 401k data and conclusions were presented in a meeting or conference, provide the raw data in the handouts. Not only does it free the presentation from slides that make the audience work too hard to decipher, but it gives you a certain amount of transparency that shows you didn’t monkey with the data to force it into the conclusion that you desired.
Richard I. Garber

Mike:

Using a pie chart on percentages that don’t really add to 100% is not good either.

If you go into Google and look up the news story at: http://www.foxnews.com/story/0,2933,505174,00.html
you can find a link to the poll data at the end: http://www.foxnews.com/projects/pdf/030509_Poll...

When you wade through that file you eventually will find those numbers are the answer to question #28, and the results in $ are as follows:

Less than 50,000 – 3%
51,000 to 75,000 – 6%
76,000 to 100,000 – 9%
101,000 to 150,000 – 9%
151,000 to 250,000 – 13%
251,000 to 500,000 – 20%
501,000 to 1,000,000 – 13%
Over 1,000,000 – 12%
Don’t know – 15%

Note there is no “other”. 15% of the people said they just plain don’t know.

A cumulative list (up to this number) would go as follows:
50,000 – 3%
75,000 – 9%
100,000 – 18%
150,000 – 27%
250,000 – 40%
500,000 – 60%
1,000,000 – 72%
More than 1,000,000 - 85%
(and 15% just don’t know)

To me it makes more sense to plot the cumulative percent in a horizontal bar chart.

I’ve said before in my blog that “Pie charts do not speak clearly; they just mumble”:
http://joyfulpublicspeaking.blogspot.com/2008/0...
In that post I referred to Professor Stephen Few’s detailed article, “Save the pies for dessert”:
http://www.perceptualedge.com/articles/visual_b...

Richard
WVMikeP

As you are I'm sure aware, I was working with the video that I linked to, which did not break it down any further. I inserted "other" because it was the only way to describe accurately that missing data from the slides they had on screen. What they showed didn't add up to 100%, but there had to be some kind of answer. I wasn't going to assume it was "don't know," "did not answer," etc. "Other" was the safest bet with regards to accuracy.

Regarding the pie vs. bar, I agree pie charts can be problematic. They are quite often overused. However, this is one case where I think it makes sense. Pie charts are the only ones that accurately display percentages with relation to the whole (100%). Bar charts are great at comparing the size of two data points, for sure.

Think about this scenario: Let's assume for a moment there are 2 data points: 25% and 75%. Your data point of interest is the one that is 25%. Now, with the numbers alone or with a bar chart, they make the value seem fairly small. It is, when compared to 75%. If they were people, they are people in the minority in some fashion. If it's an election result, it's a landslide of epic proportions. However, if your point is along the lines of 1 out of 4 people... then the visual significance of that 25% pie slice helps convey your message. Why? Well, even though it's only 25%, a pie slice that is 1/4 the whole looks fairly big. One that someone should stand up and take notice of.

A bar chart can make the minority look insignificant while the pie chart gives the minority some significance. I've seen someone use a bar chart in my 9-to-5 to make it seem like nobody's using Macs when visiting our web site. Put that same data in a pie chart and suddenly, that 1/20 or 5% seems big enough to notice and worthy of consideration.