Book Review: Dataclysm - Christian Rudder
Written by one of the founders of OkCupid, the online dating site, this book reveals how much about human psychology and behaviour can be extracted from the enormous and diverse data sets being compiled from all our digital data. For the first time in human history we have access to data sets that span billions of individuals. The title ‘dataclysm’ - a play on the word cataclysm - suggests how big a revolution this is. Each chapter reveals an intriguing data point about people's preferences and habits to dating derived from the OkCupid data set.
In the first chapter we learn what age ranges of prospective partners appear most attractive to heterosexual men and women of a given age. Averaged across all the data points, women prefer men who are the same age or a few years older then they are, however men prefer women aged 20-25 independently of their own age. The results for men run counter to modern western social expectations and this effect can be seen when looking at the age ranges of women that men of a given age message the most. The author has done an excellent job generating charts and graphics that clearly and accurately reveal what the data is telling us. If you want a lesson in presenting and explain data clearly and intelligently, this book is a great example.
Although not the author's main goal, there are some useful tips for anyone trying to improve their online dating profiles. The biggest lesson is ‘not to be meh’. The profiles that get the greatest engagement are those that are the most polarising and deviate from social norms. For example having a profile picture displaying body art, snarky expressions, or eccentric fashion. That's not to say you should engage in any of those things, but if you have an unusual taste or hobby then making that stand out in your profile would help separate you from the crowd. To phrase the principle another way: it's better to be adored by some and hated by others, than to be liked by everyone.
Rudder presents his own explanations for each of the phenomena revealed by the data which is where the subject becomes more political. Although I didn't buy all of his arguments, Rudder clearly distinguishes between his opinions and the facts, and in no way was the book espousing a political message of any sort. Overall, the book made me think about the power of data and what other interesting research questions could be investigated using these large data sets.
† In the interest of transparency the Amazon links are affiliated. If you purchase via my link, then I get a small slice in return. Feel free to use or ignore as you wish, and thank you if you decide to use them!