Book Review: The Numerati

Had a lot of plane time this week on the way too/from Beijing, so I finished off some books. Among them was The Numerati by Stephen Baker. It was a somewhat interesting read, despite the fact that it suffered from some flawed thinking.

The book makes the point that all fields are benefiting (or suffering, depending on your point of view) from a change brought on by the information age: the sudden influx of massive amounts of data. The author asserts that those that learn to master this fire hose of data and make sense of it will revolutionize their industries and beat their competitors.

The book then looks at how this availability of enormous data has affected a number of industries, talking to experts in fields including politics, marketing, security and online dating, among others.

There were a number of flaws in the book. The first of two that stood out was that the author asserts that 'the numerati' will rule the world. Those that can crunch numbers will of course have more value, but in many cases they'll just be acquired/managed/leveraged by those already in power in one way or another.

The second, and I think bigger issue, is an assumption implicit throughout the book that given a wide and deep enough sampling, all these fields are solvable or predictable. This of course isn't the case. Some problems are too noisy, some are influenced by outside factors far less dependent on past trends. And of course sometimes the link between the data and the result may non-existent (look up Phrenology if you want an example to wrap your head around ;-). The book should have spent some time looking at Wall Street quants and their collective miss on the mortgage securities bust, as an example of a field too reliant on a flawed model.

In the closing section of the book, the author acknowledges these flaws and some others as well, but after ignoring them through the rest of the book it feels like he's just doing it to offset potential criticism.

Though I found the book to have these flaws, I nonetheless found it interesting, as the anecdotes from the different fields was interesting, and overall it provoked a lot of thought on my part. I would just read it with these flaws in mind.

The Numerati