Saturday 4 July 2015

Dataclysm: Book Review

If you searched for this blog, you theoretically typed in "Dataclysm book review", sifted through the results to about page fourty or so, and decided on my blog for reasons beyond my understanding. Through that, you've done much more than you've realized; you've given Google a very minute amount of money for your search, added a tally on their search results, perhaps slightly altered the ads that will appear on your future pages, and provided one more page count on my blog (bringing it up to, in all likelihood, one). If you write about this book on Facebook, that throws you into many more metrics. Whatever race and gender you fall under, you're throwing a dime in the well for what you and yours collectively gravitate towards. You helped describe what your demographic talks about, how they speak and what's important to them - all by simply going about your daily routine. This in turn can be used to create ads that use this data from agencies that are run by people I can assume look something akin to the child of Snidely Whiplash and Mr. Burns due to their secrecy, wealth, and menacing hand motions. Now lets say you steal that line from me, post it on Facebook, and 'like' your own status as all good people do. Well, what you 'like' also says an awful lot about you as well - and don't think they (that mysterious, all-knowing 'they' that records such things) aren't paying attention. There have been programs created that judging strictly off Facebook likes that can predict within fairly reasonable accuracy whether or not you're a gay male (88%). It also has a high chance of guessing if you're black or white, Republican or Democrat (or some third party voter, or even stranger, Canadian) and it might even know if you've got a bit of a drug habit - whether you admit it or not. So if you're interested in the same sex and cocaine, be careful - your family might not know, but Facebook has got a good feeling.

All of this comes together in an umbrella term called metadata. It's all about the clicks and likes and tweets that we do on a daily basis, and how they're recorded. Dataclysm takes those numbers and shows them to us raw and uncut to demonstrate the relations between race, gender, and dating profile preferences. It's important to note that the author, Christian Rudder, is one of the founders of OkCupid, the online dating service, which provides a wealth of information on much more than you would consider at first glance. Many of the numbers and statistics come from that very site, many of which are rather expected, things you feel like you sort of knew all along but now have proof; for example, Belle & Sebastian, an indie band from Scotland, is not particularly well-received by African-Americans. I mean, if you had to bet on it, you probably would. Then it delves into some other, more surprising territory. A prime example is how men and women view each other according to attractiveness on average. Men, in spite of all the perfectly photoshopped models that abound on television, magazines, billboards, and... well, everywhere, view women's average attractiveness almost on a perfect bell-curve - in a sense meaning it's somewhat right where it should be. Women on the other hand view men much more harshly, making the average man only slightly above a 2/5. 

Rudder hits you with these facts chapter after chapter, backed up by numerous very telling graphs on each segment. The result is an enthralling page-turner that allows you to look deeply into the collective lives of large groups - which is the essence of metadata itself. It's slightly unnerving that we participate in these numbers, either unwillingly or unwittingly or both, but nevertheless they're fascinating. The range is so vast and the sample size so massive and encompassing the numbers are terribly hard to deny. It leaves precious little room for argument. On an individual level, we may not be able to use much of it for more than dinner conversation but I can guarantee you that if you read this book you certainly will wish to discuss it. It's not like it's far from the media as it is - you've almost certainly heard of the NSA and their spying tendencies as it's been at the forefront of the news for some time now. Canada just passed Bill C-51 in spite of massive opposition. This kind of information is only just dawning now, with little in the way of long-term studies to harvest from but still there is so much that can be taken from it (and argued about). Metadata is not something that will slow - we're far past that point now - and it will only become more pervasive in the years to come. Dataclysm, if anything, provides you with an early taste. 

Rudder gives his take on whether or not all the personal (yet almost undoubtedly impersonal) data collection is an evil or a blessing. We wouldn't have such wonderful free features like Twitter, Facebook, Google, Reddit and so forth if it weren't for that ad revenue that's largely fuelled by metadata. On the other hand, at what point does it feel like a flat out invasion of privacy? Like all things there's a balance somewhere in between, and it's an awfully delicate one. Dataclysm shows you the results of all that mining, for good, evil, or simple human interest. Or, just to know for sure that a very rare thing for an Asian person to type into a dating profile is "6'4". 

No comments:

Post a Comment