On Social Networks, You Are Who You Know

santosh maharshi writes "On social networks like Facebook, even if you have kept your profile very private, people can just look at your friends list and infer lots of vital information about you. Most of the social networks like Facebook and LinkedIn allow people to see your picture and your friends list as part of the open access for visitors (the article says that only 5% of Facebook users have bothered to hide their friends list). In a study titled You Are Who You Know: Inferring User Profiles in Online Social Networks (PDF), conducted by Alan Mislove of Northeastern University and his colleagues at the Max Planck Institute for Software Systems, an algorithm was tested that can accurately infer the personal attributes of Facebook users simply by looking at their friend lists. 'At Rice [University], the algorithm accurately predicted the correct dormitory, graduation year, and area of study for the many of the students. In fact, among these undergraduates, researchers found that “with as little as 20 percent of the users providing attributes we can often infer the attributes for the remaining users with over 80 percent accuracy."'"

Comments

I'm sure that marketing companies have known this

[Chrisq]

I'm sure that marketing companies have known this for years. The give-away is when they get it wrong. I get lots of adverts for cheap calls to India and for services to "send money home". I'm not Indian but most of my friends are.

Not so shocking

[calibre-not-output]

This only works assuming the public use of Facebook is ubiquitous. If only half of your friends are on the network, or if only half of them allow information about them to be publicly visible, the accuracy of the predictions will suffer greatly. This in turn means that the algorithm will more accurately predict the traits of people who have the trait of not caring about their online privacy. It's a calculation based on an assumption. In other words, bollocks.

So true

[Anonymous Coward]

I friended an old colleague of mine who has a prominent sales position at a tech firm, and was curious why he hid his friend list. So I browsed his news and watched his wall for ahwile and soon realized he just didn't want people to know he was gay. It wasn't blatent, but you could tell that a large number of people leaving messages were loudly gay, talking about gay iissues like gay marriage, etc..

Of course I never knew this whan I actually worked with him, and we litterally spent man weeks together at customer sites - although I afterwards realized that he was very good looking and never seemed to have a current girl friend, only talked about ex's. It all fit really.

So the article struck a chord with me.

What this study shows is the value of network data

[mantis2009]

For anyone who isn't clear on why Facebook and Twitter are so valuable, this study is yet another example of how much rich information is embedded in social network data. It's easy to imagine applications for pulling information out of social network data. Who would be interested in such data? Advertisers, ex-girlfriends, social researchers, police detectives, anti-terrorism, intelligence agencies... the list goes on and on. Pretty much any project with interests in the social world would benefit from social networking data. It's valuable. Why you would give away your social networking data to Facebook, Twitter, or Google for free?

Re:i'm not on facebook

[Hatta]

Henry David Thoreau said it best 150 years ago:

Just so hollow and ineffectual, for the most part, is our ordinary conversation. Surface meets surface. When our life ceases to be inward and private, conversation degenerates into mere gossip. We rarely meet a man who can tell us any news which he has not read in a newspaper, or been told by his neighbor; and, for the most part, the only difference between us and our fellow is that he has seen the newspaper, or been out to tea, and we have not. In proportion as our inward life fails, we go more constantly and desperately to the post-office. You may depend on it, that the poor fellow who walks away with the greatest number of letters, proud of his extensive correspondence, has not heard from himself this long while.

Life Without Principle, 1863

Re:What this study shows is the value of network d

[fish waffle]

It's valuable. Why you would give away your social networking data to Facebook, Twitter, or Google for free?

Actually you don't give it for free---you in fact pay them for the privilege of giving them that info (indirectly, via ads). Presumably you get something in return though.

I do agree with you on the value; personal information is a commodity, and as a commodity that is inherently mine, if you want it you should have to get my consent.

A software-licensing scheme seems like the perfect solution for personal information: you only lease my information; that doesn't give you the right to resell it, you may only use it as I explicitly direct, and I can withdraw your permissions at any time for any reason.