« Breaking up a time series | Main | Still shaken from the quake »

Comments

derek

If only the results for different years could have been arranged in a series from left to right in order of time. You could call it a sort of "time series".

Iron Man

We knew we should have added a bullseye!

More seriously, the pie chart is the right format for this kind of data since it allows us to communicate the detailed data far better than a standard time series graph using a reasonably scaled column or area chart format ever could, but a slideshow format, where the reader controls the change from frame to frame, would be a better format for presenting the data. We'll be reworking the presentation over the weekend.

We do like the "spinning wheel" metaphor though - if we really were throwing darts at minimum wage earners, there's little question of what the typical odds would be of hitting an individual within a particular age grouping, which is really the point being communicated by the chart.

Jesse Fagan

I saw this chart on Political Calculations and then unsubscribed from that blog forever. This is just a travesty. It insults the data.

Ironman

@Jesse Fagan

Oh, that was you! The data was offended that you stomped out in such a pretentious huff....

Ughhhh

Holy god this should be a histogram. The two biggest bins aren't even the same size! Whoever made this a pie chart should be fired.

Ironman

@Ughhh

Since histograms do not communicate relative percentages, they would be the wrong choice. Looks like some graph critics won't even get hired....

Wally

What is it they say about statistics and graphs - you can make them say whatever you want them to say - need I say any more?

Ironman

@Wally

Yes. May I suggest using statistics and graphs next time, because words just don't work for you....

joe

ok, I'll bite;

What makes the pie graph a better choice over something, like say this:

http://i.imgur.com/izrFC.jpg

I used the data you had. I did not clean this up at all (and used the default hideous rainbow colors, I'm sorry).

This implementation seems to show all of your data at once, with a stacked bar chart, so we can see multiple years at once, compare the percentages, and overall allow us to see the changes over time. Yea? Nay?

Ironman

@joe

Good question. The chart is fine for showing the relative percentages of each age group over time, but what it misses communicating are the large changes in the number of individuals counted within each category over the same time period.

There's room to fit that additional information on a pie chart and still be able to communicate the relative percentages for each category, but that's not achievable on a 100% stacked column chart.

It could conceivably be done on a stacked area chart, however the large change in numbers from beginning to end would make it difficult to tell that the relative percentages are fairly stable, even though the total number of individuals more than tripled.

joe

@ Ironman;

so the goal of your graph is to show that while percentages have remained constant, the number of people working at those ages has increased? I must confess that through both your writing here and the graphs themselves, this did not come across.

I will admit the stacked does not show the percentage increase in people; I did not know that was a goal.

I would then push further and suggest that the pie chart does not do this effectively either, at least not in it's current form, as you must then look at the numbers, especially over the different slides.

What's stopping this from being simply a chart with the numbers, both percentage wise and with the total number of people?

For me, I think I would use the chart I created (ok, not specifically that one, labels would be nice), and then follow it with a bar plot of the number of people for each year at each amount. Thus showing percentages with one slide, and number of people with the next. This would hit home the difference when talking about numbers and show that how you frame the question (percentage vs. total amount) causes different areas to be highlighted.

so, assuming I typed in my numbers, I would start with my first graph, then show this second graph, Illustrating the difference:

http://i.imgur.com/arqxD.jpg

(again, bad colors, etc. etc.)

I feel this conveys the dichotomy between percentages and gross numbers more efficiently. Feel free to disagree.

Ironman

@joe

I think you're on the right track in that it makes sense to develop two different graphs, although I think the grouped bar chart is too visually congested to be the second chart.

I'm wondering if a simpler column chart, showing the numerical change with a stacked bar or pie chart inset for each year, might not be a better way to go. Then, I would look at either animating it or using a slideshow format to run through the data for the number of years being presented.

What made the original animated pie chart not work very well is that it was so visually dominant (and yes, distracting). De-emphasizing it by making it a small inset against the larger backdrop showing the actual numeric changes, that problem would be minimized (literally!) Then, with the column elements showing the major changes from year-to-year, that aspect of the chart would be dominant instead, which would be much more visually appealing.

I think it would be a decent use of limited space, which is really what we're after (one relatively simple chart, communicating a lot of information), because the upper right hand corner of the graph would be otherwise unoccupied given the nature of the data, making it ideal for the inset.

I would likely keep it as a slide show that would only animate itself at the readers' choice - I think people are much happier to have full control over such graphics.

[The really scary thing is that even though the full size graphic was so annoying, it succeeded in driving increased traffic to our site from others that featured it in a smaller scale - we weren't trying to, but it appears we've accidentally validated the effectiveness of those really annoying ads involving animation at attracting clicks.]

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment

Marketing analytics and data visualization expert. Author and Speaker. Currently at Vimeo and NYU. See my full bio.

Book Blog



Link to junkcharts

Graphics design by Amanda Lee

The Read



Good Books

Keep in Touch

follow me on Twitter

Residues