Bests & Worsts Reviews from Amazon

This could be the start of a joke: Economist walks into a room, full of people who have already worked in this field for years, disregards everyone and... Only it isn't a joke. The author of this book is an economist and a NY Times author. He describes the use of Internet data, especially aggregated Google searches, for social science. In this, the book is very good. That is, if you've never heard of the idea of using these data for research, this might be a book for you. However, people have worked for several years on these data, and have much deeper, more interesting insights than Seth SD. Indeed, he seems ignorant of their work, which is a pity, because it's so much better than the work he describes. For example, Google Flu was preceded by the work of Polgreen and (independently) Eysenbach. Tracking events from Internet searches was previously done by Backstrom (2008), etc. Indeed, there is an entire academic conference (WSDM) devoted to the use of internet data. It is as if a medical doctor would write a book about the effect of drug prices on consumption, ignoring all that's known in economics. Sure, it might be interesting, but it's probably going to contain a whole lot of errors and obvious stuff. The same happens in this book. Thus, there are better books about many of the subjects described in the book, including "A billion wicked thoughts", "Dataclysm", and the recent "Crowdsourced health". Additionally, and perhaps to make the book more readable, the author makes some mistakes which will be painful for anyone with an understanding of statistics. For example, the coefficients in regression don't show causality, only correlation. In summary, if you've never heard of this idea, by all means read the book, but with caution. Otherwise, better books exist.