Paolo was kind enough to point me to Gapminder, a brilliant set of tools that help make intuitive sense of complex data, especially international development data. The folks behind the project are a group of Swedish geeks and academics who’ve committed to building free tools in collaboration with UNDP, Save the Children and the Swedish International Development Agency.
The tools that I’ve been able to play with (alas, a number of them are Windows executables…) have done a great job of giving me an intuitive sense of complex statistical issues. My favorite thus far is the set of income distribution tools that help you see distribution of resources over time in ten nations. (Most interesting, to me, is Nigeria, which from 1970 to 2000 goes from a fairly even bell-curve distribution to a curve with two maximums – a large underclass and a small middle to upperclass… making it look a lot like Brazil and the US, two nations with long-term income inequality.)
Most interesting to me, though, is a cool little manifesto buried at the end of the “Experiments” page. In it, Anna Rosling Ronnlund and Ola Rossling observe:
On a daily basis the mass media and the Internet bring us insight into peoples lives from all over the world. It has never before existed such awareness of the great multitude of cultures and traditions on our planet. Maybe this increased understanding of lives of other peple will bring us closer to each other, and serve as a platform for international collaborations. But what if the information we have about each other is not correct?
They go on to offer three reasons why this information so often is not correct: biased media coverage, a tendency to offer over-simplified explanations, and statistical ignorance. On the last, they place some of the blame firmly on the shoulders of statisticians:
Many statisticians are like musicians standing up in front of the audience showing the sheet music instead of playing it.
I’m greatly looking forward to playing with the tools these guys are building, especially Trendalyzer, their tool for animating time series data. Very, very cool that someone is taking on this issue with free software.