To increase my understanding of how Big Data works, I watched an extremely interesting TedTalks by Jennifer Golbeck. It’s called “The curly fry conundrum: why social media “likes” say more than you might think.” It’s less than 10 minutes, but they are ten, eye-opening minutes.

The curly fry example is from a finding reported in the “Proceedings of the National Academies” in 2013. The report is called “Private traits and attributes are predictable from digital records of human behavior” by Kosinski, Stillwell, and Graepel. Apparently, some really smart people “liked” a picture of curly fries and their friends “liked” the same and so-forth and so-on. And since smart people tend to be friends with other smart people, a pattern emerges that some innocuous action of liking curly fries might indicate you are smart through this association. Intelligence is only one of many traits which can be predicted with accuracy.

Golbeck’s talk is geared towards describing some of things that data scientists are able to do and some ways to move control back into the hands of users. It’s incredible to learn what small actions, such as liking curly fries or purchasing extra vitamins, can say about you in the context of millions of internet users.

She goes on to mention that if she wanted, she could easily start a company that sells the predicted attributes of candidates (good and bad) based on their information on the web to HR firms and companies that potentially want to hire the candidate. This isn’t what she advocates, though. She advocates “more science” allowing users to encrypt data or at least be aware of the risks of the actions they are taking. She advocates internet users being informed and in greater control of their own information.

The talk is enlightening. I encourage you to listen to the talk yourself. You can check it out here: link.

 

Advertisements