The Signal and the Noise: why so many predictions fail-but some don't
This was an entertaining book by Nate Silver, who I got to know during the 2010 and 2012 elections as a insightful commentator. His background in statistics and love for numbers gives a nice dose of realism to the superficial world of political commentary.
This book describes Mr. Silver's eclectic career so far and dives into several separate subjects where he beliefs his data-based analysis are useful. From climate change to the stock market, his point of view as statistician is valuable and he does a nice job explaining Bayesian logic to the general public.
The book is a little repetitive at times, and could have been 20% shorter, but this is not big deal.
Big Data, over-fitting
He is skeptical of the Big Data 'movement' which sometimes seems to imply that "if we only capture enough data, insight will follow automatically". Mr. Silver has a lot of experience with large data sets and convincingly shows the dangers of over-fitting and emphasizes that human research and insight is no substitute for large amount of data. This is a refreshing counter-argument to some of the hype in the commercial data-gathering world.
