Christian Rudder

I loved this book, and I want to say at the outset, that if you need “fodder” for starting conversations at social gatherings, this book is replete with it!

The author, Christian Rudder, is President and Co-founder of the dating site, OKCupid. He began collating and charting data from his own site, and then expanded his database by looking at comparative data from “rival” dating sites, Facebook, twitter and other social media sites. What he has found out is just amazing. The show more underlying theme is that, no matter what people claim about themselves and how they express themselves in surveys, what they actually believe and do is often quite different, but can be discovered by their online interactions. Even in the privacy of one’s own home, what one searches for in Google is revelatory.

I think my favorite aspect of this book is how it demonstrates the remarkable revolution in sociological research. It makes me gnash my teeth in regret that I did my own studies before this sort of data had become available - what fun it would have been! (Not that I didn’t have fun, in a bizarre, pedantic sort of way, but just saying….) For example, you can analyze tweets to see which people celebrate certain traditions, and how closely these mirror political borders. Using the program DOLLY (Digital Online Life and You) - to cite just one example, researchers found that the Dutch holiday of Sint Maarten is not only celebrated in the northern Netherlands, but also in Western Belgium: “the tweets reconnect old Holland to Flanders, its cultural cousin.” As the author observes: “Thus we watch an animated visualization of GPS-enabled data points, and see shadows of the Hapburgs." Just imagine, he says, if we could have tracked the tweets in Alsace-Lorraine over the years as it changed hands from German to French to German to French, with each government trying to impose its culture and language on the people. [When we traveled to that area, it was clear the mix was still trying to sort itself out!]

Other entertaining discoveries: research on Facebook has now verified that most of us are in fact connected by six degrees of separation; the majority of searches for “missed connections” are from sightings at Walmart (and most of those are in the South); when white men write essays about themselves for dating sites, the most commonly used word after “the” is “pizza”; the most antithetically used words (words most used by everyone else but least used by specific groups) for Asian men include “layed back” [spelled wrong] and ”6’4” (oddly, the second most typical phrase for Asian women is “tall for an asian”); and that the Center for Disease Control coordinates with Google to track epidemics because when people are getting sick, they search for symptoms and remedies.

Far and away the most revelatory data have to do with race and gender preference. The author explains, for instance, that a variety of indications (searches, friend connections, etc.) suggest the figure of 5% of the population being gay is pretty accurate and holds true across the states. But the number of self-reporting gays varies by the level of acceptance by states. So for example, if you see a state in which only 1.5% of respondents self-report as gay, you can probably pretty safely assume that 3.5% are in the closet. (He provides a lot of documentation to substantiate this.)

The details on race are the saddest, and show the extent to which race still is in fact a problem in the U.S. (in case you could possibly doubt it). Rudder reports data (not only from OKCupid but also DateHookup and match.com - a total of around 20 million Americans) on ratings of each group (white, black, Asian, and Hispanic) for each sex by each group, ranking the attractiveness of the other sex by race alone. Every single category and sex rates black women the lowest. Claiming to be part white elevates one’s rating substantially. Perhaps most significantly, data outside the U.S. reveal no such bias! He also talks about spikes in Google for searches like jokes about [the “n” word] that correspond precisely to peaks in Obama’s presidential campaign cycle.

In addition, as the author explains, you can find out a lot about peoples’ prejudices by watching the operation of Google’s sentence completion function. Google will fill in the most popular responses as you begin questions like, “Why do all blacks….” “Why do all gays….” “Why do husbands….,” etc.

Finally, the author includes a very thorough discussion about privacy, even bringing the Edward Snowden revelations to bear.

Discussion: I could have a couple of small quibbles. As one illustration, the author made a chart correlating the age at which a woman looks “most attractive” to a man by the age of the man. As the age of the man increases, the age of the woman by and large does not. But does that mean men find aging unattractive, or could there be a conscious or subconscious consideration that older women either might already have children (a.k.a. “baggage”) or conversely, might not be able to have children, which the man might want? [Or am I just trying to come up with reasons why aging women aren't really seen as less attractive?]

And speaking of the constraints of the data, the way questions are formulated doesn’t necessarily allow for all possible variables that might come into play. [Example from a recent Facebook “test” I took: “Do you prefer acid rock, pop, or rap? Those were all the choices; no “other”; no “none of the above”. I was forced to make a choice and provide an answer that wasn’t at all accurate.]

In other words, I sometimes want more “data” to understand the data. (Rudder says in the Afterword that he deliberately omitted statistical details to make the book more readable, because “mathematical wonkiness” wasn’t what he was trying to get across.) He does add references in the back whenever possible for further study.

Another small criticism I have is that the author presents so many arresting data findings that he sometimes goes from one to the other without full elucidation.

On the other hand, I am confident the author is aware of all of this. He acknowledges an intellectual debt to Edward Tufte, who is a (perhaps “the”) leading authority on the uses, misuses, ambiguities, and deceptiveness of data, and Rudder acknowledges:

"
…behind every number there’s a person making decisions: what to analyze, what to exclude, what frame to set around whatever pictures the numbers paint.”

As he concludes, this science of data analysis is just in its beginning stages; he is trying to give us a taste of what is already out there, and what is to come. It is aggregate data, he cautions; we still have to account for individual differences and quirks. But it sure is fascinating to find out what the numbers show about broad trends.

Evaluation: This book is full of stunning and provocative information about who we are, as well as who we want to be (but aren’t, at least not yet). Learning “sociology” has never been this fun! show less

This is a beautiful book, thoughtfully designed to illuminate insights from one large set of data, which helps us think about what can be learned from what people do and say on various vast and rapidly accumulating warehouses of words.

Interestingly enough, what people say when asked specific questions (about prejudices, for example) may be different from what they reveal in their aggregate posts. There is really no privacy concern in what is revealed here (although you could think that he show more invaded OkCupid patron privacy by even having access to it). Instead he mines his data to show interesting patterns of words used (or not used) by races, sexes, ages, and geographical areas. He shows commonalities and he shows differences— and it’s hard not to see that he is right on with his subtitle, that we really do reveal things about ourselves that we might not even consciously recognize. Different data will allow other questions to be asked, the longitudinal studies made possible by this sort of analysis will be fascinating. And he is right to get this in front of people who may think that they are in no way representative of their various identity stereotypes— they may be surprised. The book is thought provoking, the writing is wry and engaging, the mirror it holds up a clear one. show less

Christian Rudder is one of the founders of the dating site OKCupid. His degree from Harvard is in mathematics. He combines a conversational and humorously worded style with an extensive knowledge of data manipulation and presentation to deliver this treatise, which is both informative and enjoyable to read. His witty turns of phrase and common vernacular suck in the lay reader as he presents fascinating tidbits about how we utilize online services to seek a mate or communicate with friends. show more In addition to data from his vast OKCupid store, he also draws from Facebook, Google, Twitter and others to present a nuanced picture of what we can glean from the aggregate words and numbers. The chapters on race and location are particularly engaging.
Rudder's use of charts and graphs are also particularly effective. He is comparing and contrasting several sets of information and doing so in a way that is understandable and illuminating.
A person with a particular interest in race, gender and women's studies might find his conclusions particularly disturbing, as it seems that men only find young women attractive, no matter how old they are, and no one wants to date black women at all. Also, there's not really any such thing as bisexuality, as almost everyone who declares him or herself bisexual on OKCupid interacts almost exclusively with just one gender. Rudder's place here is not necessarily to moralize about or psychoanalyze his subjects, but his delivery of some of these facts seems rather glib. Whereas throughout most of the text, his humor and wit are welcome, taking the dryness out of the subject, here he seems to not be concerned about the implications of his findings.
He also seems less concerned than most people about the uses that Big Data can be put to by governments and corporations in order to control and/or monetize the citizenry. In the final chapter he does discuss the danger of all of this information and the predictive and inferential tools utilized by these entities, but considers this a "trade-off" wherein users gain access to "free" tools like Google and Facebook, and the country is more secure from terrorists.
Overall, this book is quite engaging and easy to read. The sections are well-organized and present topics of interest in a cohesive and relational way. I would recommend it to anyone interested in Big Data, but not well versed in it as of yet.
One small complaint, and probably not valid as this is an ARC, is that he mentions a tool he set up on the book's website, dataclysm.org/relationshiptest. It is supposed to tell you how good your relationship is based on the intersection of your Facebook friends with that of your partner's. Unfortunately, the site is not yet present on the Internet. After emailing the author (on August 6th) he said it would be up "later in the month". I've checked frequently and as of today, August 23rd, it's still not there. I'm a bit disappointed, because I really want to check on my relationship! Although, I do have to say that Mr. Rudder did reply to my email rather quickly, which was encouraging. The book is due to be released September 9, so I will be checking again closer to that date. show less

A few weeks ago there was a mild furor after Facebook admitted that they’d run an experiment on some users, adjusting their news feeds to include more positive or more negative items than the norm, and recording how the emotional tone of the users’ own posts changed in result. After a week or so of media discussion, Christian Rudder, a founder of the online dating site OkCupid, volunteered that his site had done something similar. He’d deliberately matched OkCupid customers with people show more rated as being poorly compatible and monitored reports about the resulting in-person dates.

He accepted interviews to explain himself, and in them he was absolutely unapologetic. Not only does he not see an ethical problem in defying user’s expectations about the site’s behavior, he says that an online company that does not engage in this kind of experimentation is irresponsible, because it is foregoing opportunities to learn how to improve its customer-aiding algorithms. He is convinced that every online company engages in similar tests, and asserts that they do as a fact.

This is the context into which Rudder’s new book, Dataclysm, is released. In it, Rudder reveals himself as an enthusiast for both data and people—at least when considered in the abstract. His thesis is that the Internet has made it possible, for the first time in history, to obtain (almost as a side effect) massive amounts of data about what people are really like—how they behave and what they think in secret. The fourteen chapters are based around broad themes and contain many interesting maps, graphs, and charts, startlingly illustrating what the data tells us. The charts are like salted peanuts. It’s hard to find the will to stop consuming them.

Rudder presents the data in an engagingly enthusiastic, casual way. At times he achieves a little poetry, such as when discussing a map showing the density of reactions on Twitter to an earthquake, overlaying a map showing the epicenter: “Here we see contours of surprise laid over the shifting earth.” At other times he seems clueless, even crude. This is nowhere more true than in the first chapter, which is devoted to the hoary truth that young women remain equally attractive to men throughout men’s lives. Rudder treats the depressing data with a glib tone, and troublingly, doesn’t pause to consider that when older men date older women, an explanation may be that older men have learned that sexual attraction is only one ingredient in a healthy relationship. He seems to prefer the explanation than men are simply hypocrites, pretending to desire that which they do not.

It’s a shame that he leads with this chapter, because the rest are better. Did you know that you’ll tend to date more frequently and successfully if people either love you or hate you than if everybody just thinks you’re kind of cute? That people in Mississippi bathe more often than people in any other state? And what does it mean that women are more “race-loyal” than men, preferring more strongly than men do to date people of the same race?

In addition to presenting interesting data, Rudder discusses what kinds of data about your life is being collected by various types of sites and agencies, and demonstrates the surprising ways they can use the information. Facebook and the NSA, of course, are part of the discussion. In fact, I’d go so far as to say that if you have any interest at all in the data that is being collected and shared about you, you’ll want to read Chapter 14. Rudder isn’t completely deaf to the implications of data mining, including the ethical implications. He just thinks he falls on the right side of the line. Others disagree. show less