Skip to content
Surprising Science

Do Data Miners Know You Better than You Do?

A few simple concepts power today’s powerful data mining algorithms, allowing companies and governments to gather data about who you are, what you do and what you might do. 

What’s the Latest Development?

It’s nearly tax time, and if your return looks anomalous with respect to the millions of others filed, the IRS might take a closer look. That’s just one of the ways in which powerful computer algorithms can extract personal data over a wide range of information to give companies and governments a clear vision of who you are—and what you might do in the future. Another technique, called association learning, is currently used by Amazon to predict what you might buy in the future based on what other people bought after buying the same product you purchased. 

What’s the Big Idea?

A digital signature is created and stored each time an interaction occurs over the Internet, meaning that the things you say and buy online are part of a massive data bank open to analysis. Regression is a technique that compares a company’s predictions about your behavior to how you have actually behaved, allowing them to modify how they make their predictions, in hopes of improving them. Facebook, for example, might use regression to design their site in a way that encourages more interaction, something they have presumably done with Timeline. Based on initial user response, we might ask if Facebook knows us better than we know ourselves?

Photo credit:


Up Next