Yesterday while the net was down at home, I did some real reading -- G. H. Hardy's A Mathematician's Apology. It's a good entry to how a serious creative mathematician looks at his subject, with some interesting general philosophy. I don't agree with everything he says; but he says it well, so I have some work to do to crystallize exactly how and why we disagree. Fun!
Not a recent reading as a classic that I've recently mentioned elsewhere. It's a very good book to read if you read ... pretty much anything. How to Lie With Statistics by Darrell Huff, illustrations by Irving Geis. The recent mention was prompted by someone who wanted to illustrate that there's no trend or not much -- by showing a graph where the vertical axis was far larger than the data range. This is a technique for lying straight out of the book. Once you've read it (it is old, 1954, so some current methods may not be mentioned, you should be able to recognize other methods of distortion.