I had a treat over Christmas avoiding Twitter and reading pre-releases of books about maths. The joys! Two particular pearls are about to be released.
The Art of Statistics (US) (UK) by Sir David Spiegelhalter should be self-recommending, but this is a really first class introduction to the power of statistics. David starts with some basics (categories, proportions, visualisation) but by the end of the book has covered big data analytics, confidence intervals, Bayesian statistics and much else. It’s a remarkably accessible read, full of powerful examples, but covers technical ground too, where appropriate. I can’t think of a better starting point for someone who wants to become a statistician or to use statistics in any professional way, and it covers most of what the lay-person would need. Bravo!
Humble Pi (US) (UK) by Matt Parker is a very funny collection of tales of mathematical, programming or engineering errors, generally with non-fatal consequences, although there are a few billion dollars lost here and there. Matt smuggles in a great deal of wisdom and geeky detail – for example, how to produce a rounding error when asking Excel to subtract 0.4 and 0.1 from 0.5. I loved the book.
I’ve also just caught up with the existence of Is That A Big Number? (US) (UK) by Andrew Elliott, which offers much wisdom for putting numbers into perspective by visualising, estimating or comparing them. One idea I particularly liked was the “landmark number” (for example: a book is about 100,000 words long; it’s a 3000 miles or 5,000km drive from Boston to Seattle) – having a few of these numbers in your head or at your fingertips for comparative purposes is much to be recommended.
Next up, Invisible Women (US) (UK) by Caroline Criado Perez, about the way the data we gather often omits or short-changes women. An important topic and the book is getting good reviews. I’ll report back.
UPDATE Friday 1 March – having read the first 100 pages of Invisible Women I can report that it’s an excellent, powerful and thought-provoking book about the way our lives revolve around the assumption that “man” is the default and “woman” the weird edge-case. Examples from interior design (Le Corbusier designed the proportions of his interiors around average men, dooming the average woman never to be able to reach the top shelf) and snow-sweeping (men are more likely to drive, women are more likely to walk: do we clear the roads first, or the sidewalks?). Despite the subtitle (“Exposing data bias”) there is not much yet about data bias but still time for the book to scratch that particular nerdy itch.
ANOTHER UPDATE Tuesday 5 March – quite a lot of v. interesting stuff in the second half of Invisible Women about subtle (and less subtle) biases in the data we collect. For example – gathering data on household income (rather than individual income) isn’t a crazy thing to do, but it does obscure any question of who in the household is earning and / or controlling the cash. I’ll be writing about the blind-spots in our data for the FT this weekend.