Wednesday, 25 May 2011

I'm getting a little worried about myself - I seem to be developing an interest in statistics.

I am currently reading two stats books with relish:

Edward Tufte "The Visual Display of Quantitative Information"

Peter Dalgaard "Introductory Statistics with R'

The Tufte especially I'm finding well nigh unputdownable. A beautiful argument embodied in a beautiful book.

It all started with an inspirational little stats course at UCL earlier this year, which began with the story of the Challenger shuttle disaster in 1986. At a meeting the night before the launch all the data needed to predict the failure of the O-rings was presented by the engineers to the managers, but in such an opaque way that the message failed to get across and the flight went ahead anyway.

So good stats can save lives, folks!

Later on the lecturer introduced some ideas from Tufte, and showed us some nice graphs done with R.

So now I'm a bit hooked. Playing with R in my spare time. And thinking about super-clear ways to display the data I hope to get (one day!) from my experiment. And today I was even asked about stats by some of my lab-mates. Not sure if I answered correctly, but need to be careful here - getting a reputation.