Who We Are (When We Think No One's Looking))] [Author

This is a most interesting book. Its insights are presented with droll wit. Don't, however, buy the kindle edition. The data is presented in a set of very interesting graphs, none of which you will be able to make out entirely on a paper-white device. They are too small and the graphic is too faint. That is vexing. The charts are at least a quarter of the value of the book, so by paying for hard-copy you'll more than get your money's worth.

Interesting facts on how people behave on the internet when they think they are anonymous. Answers on why copy and past works for a lot of things, how being normal relates to being strange/extreme when it comes to getting attention.

Fun and entertaining, I loved this book!! As a graduate student of economics and data science, I found this book to be intriguing and eye-opening to a whole new world of analysis that I didn't even know existed. Bravo!

As a statistics nerd, I really appreciate the quality of data presented in this book, the variety of algorithms used, but overall this book doesn't offer much insight into data as an analyst would expect. Yes, Belle and Sebastian is the least black band in the world, but what does it mean? Can you just take this single fact out to make an analysis? The author points out how he purposefully leaves the data up for interpretation, and I appreciate an opportunity to look at the statistical data and draw conclusions myself, but I think in the end the author just tried to make a point how many anecdotical conclusions can be drawn even from the anonymous big data. Yes, it's possible and it's true, but I don't find it scary, I find it fascinating. The real insight is what this book is missing.

Dataclysm covers an interesting topic - big data - in a disappointing fashion. Rudder demonstrates how big data can be used in creative and innovative ways, but the focus on OK Cupid diminishes the effect. The results Rudder presents about OK Cupid data are mildly interesting. However, these results have already been highly publicized on Rudder's blog and in the media. You are unlikely to find any new information about human attraction and dating in this book. Rudder does introduce some data from Google, but it is underemployed throughout the book.

There are interesting ideas raised by Rudder - racial cleavages in particular - but the majority of Dataclysm is about the banality of dating behavior. In addition, Rudder's methods with this OK Cupid data are occasionally questionable, albeit original. For instance, Rudder "demonstrates" the harsh differences between different races and gender by using an algorithm that detects words that are supposedly characteristic of the demographic group in question. The mechanics of the algorithm, however, are to detect words used by the demographic group, but not used by any other group. When these words are presented, there is no information on how common they are among the group in question. If these words constitute a high percentage of the group's language, this algorithm seems reasonable. However, it seems more likely that the words are rarely used within this demographic group. Thus, the algorithm is constructed to show large demographic differences, even if the differences are in reality miniscule.

Finally, Rudder writes with a snarky attitude that rubs readers the wrong way - even those readers who agree with his sentiments. These comments indicate that Rudder is overly confident in the importance in his small findings. Perhaps this is why he has overlooked some of the flaws in his methods.

Recommendation: if you are looking for inspiration on using big data in creative ways, skip part 1 and skim parts 2 and 3. If you are interested in human attraction, skip the book and read OK Cupid's blog; you'll save money and time.

This book discusses the author's observations from OKCupid, a dating site, with some reinforcements from other web sources, such as Twitter and Google.

The work on OKCupid is quite interesting: who wants to date who, males and females' preferences, etc. It was great in Rudder's blog, but I felt it didn't transfer well enough to a book. Rudder tries to enhance his data with sociological information and his views, but these are not altogether very convincing. He is most in his element discussing OKCupid.

It is also curious that the author hasn't published any of these data in academic journals. This makes it (in my view) a less valid observation.

Finally, there are a few minor mistakes in the book. For example, it is doubtful that there is research on civil unrest in Egyptian cities bordering Israel, because there are so few of the latter. Also, when Rudder discusses words, but shows phrases, it detracts from the message because it is inaccurate.

Wow. Thanks for the tip. This is a great resource for psychologists interested in research. It simply peels away the accumulated layers of 'varnish' we paint on day after day as a defense from perceived attackers.

As Christian notes, it is very important to have intuitions confirmed with great data. Unfortunately, there aren't that many surprises, and some aren't pleasant as they reveal us as more banal than we'd ever like to admit.

This book is fascinating, insightful, and hilarious. It's sort of a survey of the types of things that are possible to see about the way people behave and who they are on a large scale at this moment in place and time. It's using Big Data to look at people's behaviors and preferences and so forth not to try to sell them something or to see if they're terrorists but just to see who we are as people, to help us see things about ourselves, and I think hopefully to help us ask ourselves some tough questions. It's not comprehensive, because considering the huge trove of data and the complexity of humans, that would be nearly impossible, but the areas the author chooses to examine are thoughtful, poignant, and at times downright surprising. A lot of things seem to just show things that we might have thought intuitively, but that didn't make it less fascinating to see the behavioral statistics showing it to be true. It's easy to be like, "men are like this and women are like this" or "black people are like this and white people are like this" or whatever, but to see some of this stuff borne out so starkly in the data, I thought it was incredible. And the writing is highly readable. Somehow he turns a book about data into an enjoyable page-turner that I didn't want to put down. Much of the book is laugh-out-loud funny, while some sections are sobering, bordering on depressing. But none of it was dry or boring. And I'll probably read it again.

In short, a mildly amusing, somewhat inaccurate popular science portrayal of BigData from the viewpoint of a former OKCupid head honcho - who is not all that good at statistics or experimental design.
----------
The Good:
The book is a lively and occasionally interesting work which you will find of interest if you use OK Cupid a lot or don't know anything about BigData. The graphs in the book are very pleasing to the eye.

The Bad:

This person thinks Wikipedia is a reputable source for, well, everything. Sources are misattributed. Some facts are just plain wrong - like embeddedness' origin (it's Polanyi). A lot of the statistics involve sweeping and bizarre assumptions (sentences have no impact on language) or incorrect extrapolations of information (Twitter is far superior to Shakespeare).

Profanity is used needlessly (this would otherwise be perfectly appropriate for young teens). Last, but certainly not least, sources are unreliable (Wikipedia is used repeatedly). The book is disorganized and desperately needs some chapters juggled around to be a more cohesive entity. I blame the editor as much as the author.

Overall, very readable, but thoroughly mediocre and terribly sourced.
-------------

Mediocre. 2.5/5

A Disappointment

I am a user of OKCupid, and have used other dating sites in the past, yet I am still single. My intention, upon reading this book, was to learn ways in which I could improve the effectiveness of my search for a soul mate. This book is practically useless for that purpose.

The author, Christian Rudder is married, so the topic of finding one's soul mate isn't very interesting to him, and he pays little attention to it in this book. Politically, he is a liberal, and as such wants to talk about practically nothing other than race and sexual preference, and the things he wants to say about them are entirely predictable given his political bias.

He claims to have data on how many sexual partners people have had. I've answered about 900 questions on OKCupid, but none asked me my number of sexual partners. And he concludes that male gays have about the same number of sexual partners as anybody else. This wildly conflicts with what I've heard from male gays themselves, and from what can be discerned from rates of transmission of STD's (not just HIV) among different groups. It's possible that the most promiscuous people meet each other in bars and bath houses and don't show up on OKCupid.

About race, he notes that black women and Asian men are the least sought-after quantities in the dating marketplace, something I knew before I started this book. Black men also fare pretty badly, which surprised me, because I had the impression that women of all races found them physically attractive. I was also shocked that Asian women weren't doing exceptionally well, but that may be just my own bias -- I am white, and have dated lots of Asian women.
Rudder speculates that given that so much discrimination on the basis of race is going on when people choose life partners, this indicates that there is probably a lot of discrimination when processing mortgage loan applications, as well.
This doesn't follow. For example, blondes probably do very well in the dating market, but that doesn't establish that they have an easier time getting home loans.
There was a famous study done in Boston around 1990 that indicated that there was a lot of racial discrimination by banks when giving home loans, and after Clinton got elected, regulators descended on the banks and made their lives miserable over it. But it turned out the study was flawed, taking only debt-to-income-level into account. When other factors such as employment history and credit record were taken into account, no significant discrimination was found.
People processing home loans are not making a decision about whether they will have to deal with someone every day. It's a numbers game -- your employer, the bank, gives you a formula, and you go by the numbers. Whether the person gets the loan or not is of little consequence to the loan processor, what matters to their job is: did they get the numbers right?