Food art
Seen at Starbucks

Some readings

Here are some good reading from other sites:

Harvard's Social Science Statistics blog reported on a talk by Amanda Cox of the New York Times graphics team, without whom this blog would lose a big source of inspiration.

Over at the Gelman blog, Aleks and Andrew discussed map-making, and as we often say here, picking the right things to plot is equally, or even more, important than what the graph looks like. 

Some time ago, Andrew used a variety of graphs to look at the industry of predicting election results.

Dan Goldstein showed a graph that supposedly demonstrated the fallacy that we all think we are above average in the income distribution.   I find it interesting in a different way: it showed that people who are very wealthy or very poor know their "place" so to speak but people in the middle have a difficult time knowing where they stand, which seems logical to me.  Dan has previously written about graphs

What other blogs are saying about us:

Infographics News: "You can agree or not with the author's point of view, but it's a good place to visit before making our next pie."

Opposed Systems Design: "This chart is just crying out for a Junk Charts treatment."

Decision Science News: they mentioned us in this post about Minard's Napoleon map, which Tufte considers the best graph ever.

Happy holidays!  As usual, I've updated my Wish List (link on the right): help build my library!


Feed You can follow this conversation by subscribing to the comment feed for this post.


people who are very wealthy or very poor know their "place"

They could hardly not, given that they're at the left and right walls. Also, at the stratospherically high incomes, percentile is little better than decile as a precise measure of relative wealth.

I'd be interested to see the same graph re-scaled using a probability scale, such that zero and 100% are at infinity. Will the rich and poor look as accurate as they do on the more bounded linear scale?


I agree looking at percentiles would be meaningful. But because of the upper/lower bounds, the variance at the extremes has got to be smaller than that in the middle of the distribution.

The comments to this entry are closed.