## terrible graph of the day

A truly terrible graph in Le Monde about overweight and obesity in the EU countries (and Switzerland). The circle presentation makes no logical sense. Countries are ordered by 2030 overweight percentages, which implies the order differs for men and women. (With a neat sexist differentiation between male and female figures.)  The allocation of the (2010) grey bar to its country is unclear (left or right?). And there is no uncertain associated with the 2030 predictions. There is no message coming out of the graph, like the massive explosion in the obesity and overweight percentages in EU countries. Now, given that the data is available for women and men, ‘Og’s readers should feel free to send me alternative representations!

### 6 Responses to “terrible graph of the day”

1. It’s not easy to try representing a value x 2 categories x 2 sexes x 2 years :)

A line graph could work if there were less overplotting on the y-axis:

2. Do you think ‘R ‘ can create this ?

• You can come very close to representing the same kind of information with ggplot2: here’s an example.

And since ggplot2 supports polar coordinates, you should be able to reproduce the original graph. It’s just not a good idea at all: polar coordinates are impossible to read properly.

3. Circular graphics are often used when there is an abundance of categories where the graph creators show all the categories (e.g., in phylogenetics: http://itol.embl.de/), so I understand the reasoning behind it. I understand that they were going to try to represent the data by country, but then they had four different ways of ranking the graphs: combinations of sex and year. The uncertainty thing also bothers me, too. To answer your question as to how to make it different, I’d make a simple scatterplot with the 2010 weights on the abscissa and the 2030 weights on the ordinate axis. There’d be a 1:1 line showing no change under the points, and the symbols different for males and females (the Mars and Venus symbols). Because males weigh more than females, they’d be bimodally distributed, so there’d be about n-countries symbols for each sex in the scatter cloud, which isn’t too crowded (there’s also a descent amount of variance within sex from eyeballing this graph). What would be fancy about it–and what would preserve the information content of the above graph–would be that it’d be interactive and a scroll-over would pop-up the name of the country. Thanks for the post!

4. biostatmatt Says:

The circular presentation, radiating lines and color draw attention and give an aesthetic that is totally unrelated to the information content, which could be important on a busy newspaper page.

I sometimes enjoy the challenge of deciphering graphs like this, much like the puzzles that you write about, when I’m reading for recreation.