Okay, unlike the last big data book I read, this WAS the big data book I wanted to read. While it does describe the mathematics used in the different ways big-data affects society, it does show how big data, how using small bits of innocent data can reveal surprising truths about who we are and what we really think.
If you start with the assumption that everybody lies on surveys (and this isn't a bad assumption, people mess with surveys all the time, to make themselves look good, to make someone else look bad, to withhold information for ulterior reasons, or really just to f--- with the survey data), then the fundamental data from which theories and beliefs develop is wrong.
But put people in a place where they believe their thoughts are anonymous, truly believe the data they provide will not be traced back to them, then said people become more open and, well, more honest with what they are thinking.
Which is where Stephens-Davidowitz's data research idea came from, to use Google search data as a research source, and where many assumptions about what people are thinking can be debunked.
And the results are fascinating.
Stephens-Davidowitz provides a number of examples of "here's common knowledge, everyone knows this," and shows where, with big data, the "knowledge" is wrong. Either we aren't the same as when the knowledge was first determined, or it was declared as true based on something unknown, or simply accepted as true based on some voice of authority. Regardless of the source, the actual data, the actual numbers, show a different story, and that is the fascinating part.