« Two lines dropping: further reading | Main | Colors and scales break the seam between charts »


Disinterested Observer

Wouldn't such data be better represented with a Lorenz curve?


A real Lorenz curve can't be derived from the published data precisely because the outer brackets were defined as "less than X" and "more than Y". We can't then recover the total income for those brackets, and for obvious reasons, not being able to get the total income for the top bracket is highly problematic.

I first tried the next best thing, which is to keep the income brackets as the horizontal axis, and plotting the cumulative proportion of population as the vertical axis. Now, readers are forced to examine the vertical distance between two curves in order to understand which part of the distribution diverges from the national average. Thus, I came to this chart where the vertical distances are plotted directly.

Eric Obermühlner

While going through the full history of this great blog I stumbled over this entry and to my suprise I find that I don't agree with you.

While the chart is visually beautyful I find several details quite annoying.

The black "NATIONAL" bar looks more like a subtitle to me. Took a while to realize it was referring to the horizontal stacked bar below it.
I agree that the national bar should be vertical and shown next to the others.

Sorting by the size of the richest group would probably greatly enhance the readability.

I found the bottom line especially confusing since I could not figure out what was encoded here (it looks too dark in many cases to be the mean - but I could be wrong). Why not label it?

The comments to this entry are closed.


Link to Principal Analytics Prep

See our curriculum, instructors. Apply.
Marketing analytics and data visualization expert. Author and Speaker. 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