Data Visualisation: the Good, the Bad and the Ugly (3)

How do you sort through your data and show what’s relevant?

Cookies! This time, Puzzle #3 is about Girl Scouts selling cookies and comparing scores. We’ll see how showing too much data can be dangerous and how our brains can be tricked by graphic elements even if, in truth, they don’t make much sense.

To learn more about the GBUDV project, read the introductory article.

The story

Since 1917, selling cookies has been a way for the Girl Scouts of the USA to raise funds. They offer different kinds of cookies and the Girl Scout Cookies profits are used for activities, events, maintenance, administrative costs and so on.

The Little Squirrels’ Troop super brand!

Our little troop of 11 girls, the Little Squirrels, sells its own Delicious Cookies. Now, they’re looking at the results to decide who’s the best saler. Each one has been assigned a random number from 1 to 11.

The Good’s message:

Our 11 girls have varying degrees of success. Camille and Jane are clearly ahead while Eleven has sadly not been very efficient.

The Bad’s message:

Sure. Yeah.

The Ugly’s message:

Wow, sales are so going up: see this line in the middle that rises steadily?… What do you mean it’s got nothing to do with it?

The puzzle

Why is the Good good?

In this example, we simply want to compare scores for one value. This size-comparison visualisation lets you match each result to each girl with ease, plus your instinct tells you the bigger the circle the bigger the sale so you know who did well. We even have a different symbol to highlight the high value: who could ask for more?

(Only downside: the limited colour palette assigns the same tone to several girls, which can disturb the audience… always think of how large you need your colour chart to be!)

Why is the Bad bad?

Besides outputting simple figures as proportions which has absolutely no advantage and only loses the reader, this radial chart makes another mistake: proportions are given according to the highest value (90). So you don’t have percentage but ‘per-ninety-age’ which is really weird. Just look at the results for Hope: the circle does stretch half the total distance but she didn’t sell 50 cookies, only 45.

The title is the only indication of this strange range of values.

The number of colours has even worse consequences than in the Good: since you have to go back and forth from the legend to the graph, your eye has trouble linking the right circle to the right girl. And even so, by the time you did the association, you probably forgot what you had painfully deducted from the arc length, right?

Why is the Ugly ugly?

Ah, the line that goes up! That’s what you want to see on any sales report, am I right? The only thing is: this one is completely dumb and false.

You have three tricks here:

  • we are introducing into the plot an arbitrary variable that has no link whatsoever with the rest of the data: the girl’s random number; while having a supplementary criterion is not inherently a problem, we are deceiving the audience by presenting it as an essential part of the study. Since it gives the points’ height – which instinctively determines their importance in our minds – and even gives a trendline, it must be important…
  • the axis are intentionally reversed! This fools us because we naturally assume it is the vertical distance to zero that gives the most information, and not the horizontal distance: who would consider a point further on the right represents a higher sale? At first, you will rather examine the y-positions in the graph to determine the best results.

A more honest representation of the situation would be the one below where we switched back the axis to a ‘normal’ order – where you immediately see Camille’s and Jane’s performance:

  • this Ugly knows it could be found out quickly… so it plays with the data to convince us it’s displaying the truth.

Take a look at the point on the right: this is Jane’s result. We know she did well and figures support this heavily. So a misleading speaker could begin by asking the audience to check for themselves that Jane, who’s high in the chart, sold the most. This would unconsciously strengthen the link we instinctively made between y-positions and sales results. Thus when the speaker goes on and tells you that Kate did even better – see, she’s higher in the plot -, you probably will agree… even though she only sold 30 cookies!

Of course, we also made sure we had more colours so there is no doubt which point is assigned to which girl: if you had to think about it, you might see there’s a problem.

A peek at the next puzzle…

For the final puzzle, we’ll stay in the food industry and talk about the famous and completely real ChocoChoc’s Factory that sells candy and chocolate all around the world! 🙂

Leave a Reply

Your email address will not be published.