Your Facebook Likes Reveal Your Age, Gender and Ethnicity?

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
According to a new study, researchers are able to predict personal information about a group of people based solely on the patterns of likes on Facebook.

The researchers found, for example, that "Likes" for Austin, Texas; "Big Momma" movies; and the statement "Relationships Should Be Between Two People Not the Whole Universe" were among a set of 10 choices that, combined, predicted drug use. Meanwhile, "Likes" for swimming, chocolate-chip cookie-dough ice cream and "Sliding On Floors with Your Socks On" were part of a pattern predicting that a person didn't use drugs.
 
Apparently drug users are more likely to want monogamous relationships?

Does that lso sugest that non-drug users are more likely to sleep around? :p
 
It seriously bothers me that useless studies like this actually get funded.

They aren't always as useless as they may sound.

Regression analysis is a very powerful tool, but before you set out to analyze your data, you don't really know exactly what patterns you are going to find. Some of them may be very useful, others, well, not so much.

Target is VERY effective at using regression analysis on their customers history. In some cases they have been too good at it, figuring out that their customers are pregnant and sending them deals and coupons on baby stuff before they even know they are pregnant themselves, which can come across somewhat creepy :p

There was one case of parental outrage when Target started sending a teenage girl baby deals before her parents (and possibly she, herself) were aware.

This article in Forbes goes over it. It's very interesting.

I have been known to shop at Target a bit. I wonder what their regression analysis tells them about me :p
 
I would think cookie dough ice cream would be a sign of drug use.

Anyway, as a 33 year old white male that doesn't use Facebook, I clearly have a concrete bunker and don't trust banks.
 
I'm pretty sure this thread will be filled with attempts to attack, dismiss, or debunk by Facebook users who feel threatened by the fact that they can be picked apart by the things they decided to "Like" as a means of putting their own minds at ease over their addictions.
 
I would think cookie dough ice cream would be a sign of drug use.

Anyway, as a 33 year old white male that doesn't use Facebook, I clearly have a concrete bunker and don't trust banks.

Well, that's a common misconception about regression analysis (and statistics in general).

Based on patterns statistics can predict certain behaviors or attributes on a category of people, but it can't say anything for certain about the individual.

Take the Target example above. Let's say their statistics predict that individuals with certain shopping habits predict a 70% chance of pregnancy. Makes sense to target those people with baby coupons. They are going to be wrong about 30% of the people they send the coupons to, but it is still a way more targeted ad campaign than just carpet bombing everyone with baby ads.

Same thing with big brother.

Statistical regression and behavioral analysis can narrow down the field of suspects of devious behavior, but it will never show an individual to be guilty or not. It just helps authorities look more closely at those who are more likely to be guilty of something, and then in the process, hopefully finding the evidence that prevents the next 9-11, rather than spreading their thin resources over millions and millions of people.
 
I'm not the only one who doesn't understand the value of liking right?

I mean, why the fuck would i want to go around liking shit?
 
I'm not the only one who doesn't understand the value of liking right?

I mean, why the fuck would i want to go around liking shit?

People feel they are expressing who they are, and their personality by prominently showing off what they "like" on their profile pages.

What they don't realize is that once you "like" something, whoever owns the page you like, has access to all your data.

Same goes for once you give permissions to a meme or a little facebook game.

Also, unless your friends have explicitly turned it off (it's on my default) they usually share all their friends data when they do this as well.

In fact, many (most?) "like" pages, games and meme's are really not into it for the business of games or memes and the like, but rather these are just clever ruses in order to harvest personal data they can use to their advantage.

Yes, I am on Facebook (Albeit, reluctantly. I find it difficult to keep up with my friends without it these days, since everyone communicates and plans events via Facebook), but I have the applications platform disabled, and I don't do "likes" due to the privacy concerns.
 
Zarathustra[H];1039695304 said:
since everyone communicates and plans events via Facebook),
This is why i use facebook too.

And thanks for your informative post (seriously). I always found it funny how someone of my friends go through and like every single comment on something, it's just weird as hell to me. But i'm glad to know about the security concerns of doing that. I'm going to go make sure my application shit is turned off too.
 
Zarathustra[H];1039695304 said:
People feel they are expressing who they are, and their personality by prominently showing off what they "like" on their profile pages.

What they don't realize is that once you "like" something, whoever owns the page you like, has access to all your data.

Same goes for once you give permissions to a meme or a little facebook game.

Also, unless your friends have explicitly turned it off (it's on my default) they usually share all their friends data when they do this as well.

In fact, many (most?) "like" pages, games and meme's are really not into it for the business of games or memes and the like, but rather these are just clever ruses in order to harvest personal data they can use to their advantage.

Yes, I am on Facebook (Albeit, reluctantly. I find it difficult to keep up with my friends without it these days, since everyone communicates and plans events via Facebook), but I have the applications platform disabled, and I don't do "likes" due to the privacy concerns.

First off I'd like to say QFT to all your previous posts. Now this "liking" explanation isn't 100% clear. You only allow pages access to your public data by liking the actual page. Liking comments, photos, or other stupid shit does not give the source of said content access.

But yea, people like to show of likes I guess or maybe they like a page so they can keep up with it. I liked the Battlefield page and I enjoy their posts that show up on my news feed. Now reading the comments might lower your GPA but, every website suffers from that to some extent. I also think people like stuff (not pages) because its easier to click that icon once, than actually typing out that you like it and optionally, why.
 
First off I'd like to say QFT to all your previous posts. Now this "liking" explanation isn't 100% clear. You only allow pages access to your public data by liking the actual page. Liking comments, photos, or other stupid shit does not give the source of said content access.

Agreed. I should have been more clear. Liking a friends post, picture or comment is not a problem.

Liking a company, fan page or website does though. (as well as the application platform stuff)
 
Zarathustra[H];1039695285 said:
Based on patterns statistics can predict certain behaviors or attributes on a category of people, but it can't say anything for certain about the individual.

It would be WAY off for me. I'm 27 currently, but I've been told by various people for years now that a lot of my tastes in music and some other stuff make me seem a LOT older than I actually am.
 
I'm not sure what there is about this that actually surprises anyone.

Some have question funding. Well, the data is in Facebook and it's all tallied by computer. All it takes is someone asking the question. There is no conspiracy or anything else. This is how this stuff works. As a group of "tech savy" individuals....people in this forum should be perfectly able to understand this. Things like this really should go without saying (shouldn't it?).

As for connecting Austin to drugs....I'm not surprised at that either.

All kinds of accurate connections could be made with things like this.
 
I am not surprised at seeing a study like this. After all, we have had studies that determined men and women are different. :eek: :D
 
Zarathustra[H];1039695236 said:
They aren't always as useless as they may sound.

Regression analysis is a very powerful tool, but before you set out to analyze your data, you don't really know exactly what patterns you are going to find. Some of them may be very useful, others, well, not so much.

Target is VERY effective at using regression analysis on their customers history. In some cases they have been too good at it, figuring out that their customers are pregnant and sending them deals and coupons on baby stuff before they even know they are pregnant themselves, which can come across somewhat creepy :p

There was one case of parental outrage when Target started sending a teenage girl baby deals before her parents (and possibly she, herself) were aware.

This article in Forbes goes over it. It's very interesting.

I have been known to shop at Target a bit. I wonder what their regression analysis tells them about me :p

There is a case study in Data Mining regarding how using data mining technics Walmart was able to find a patern that allowed them to move beer six-packs near diapers and had a 30% improvement on sales. All because they found out a pattern where men buying diapers also bought beer.
 
There is a case study in Data Mining regarding how using data mining technics Walmart was able to find a patern that allowed them to move beer six-packs near diapers and had a 30% improvement on sales. All because they found out a pattern where men buying diapers also bought beer.

that's awesome.
 
Zarathustra[H];1039695304 said:
Yes, I am on Facebook (Albeit, reluctantly. I find it difficult to keep up with my friends without it these days, since everyone communicates and plans events via Facebook), but I have the applications platform disabled, and I don't do "likes" due to the privacy concerns.

Friend don't make friends use facebook :)
 
There is a case study in Data Mining regarding how using data mining technics Walmart was able to find a patern that allowed them to move beer six-packs near diapers and had a 30% improvement on sales. All because they found out a pattern where men buying diapers also bought beer.

There's actually correlation there. Regardless of what data shows, liking Austin, TX has absolutely NOTHING to do with the chances of you being a drug user. Or any other stupid shit they might just happen to work with some group of people. No matter how fancy your algorithm, you can't add correlation to unrelated things. Ever.

The only reason stupid crap like this might be funded is for use in advertising, or some Minority Report type shit. Neither of which do I want (in the case of advertising, I'd rather it NOT be personalized - advertisers should not get the privilege of knowing ANYTHING about me unless I actively choose to give them information).
 
Zarathustra[H];1039695042 said:
Apparently drug users are more likely to want monogamous relationships?

Other way around. Monogamous relationships are more likely to cause drug use ;) (ok I'm actually for monogamy, but I still think it's funny).
 
All this may mean I'm a 12 yr old transgender martian :(
Would explain a lot though.
 
It would be WAY off for me. I'm 27 currently, but I've been told by various people for years now that a lot of my tastes in music and some other stuff make me seem a LOT older than I actually am.

But if they have access to your music listening habits (like through Last.fm or Spotify) then maybe they can predict your personality based on that? :p
 
Back
Top