How dangerous is Big Data?
​

Now, anyone who did not live on the moon the last five years is familiar with the term Big Data. This business softwareterm means and that all that we do - on the Internet or "offline" - leaves a digital trail. Purchase by credit card; you can ask Google, walking with a smartphone in your pocket Like everyone in the social network: all that is saved. For a long time, no one could understand what good could come in handy this data - except in cases when the advertisement of funds from hypertension pops up in the Facebook feed because recently we searched Google for "how to lower the pressure". It was not clear, what means Big Data for mankind - a great danger or a great achievement?But on November 9, 2016, we found the answer to that. After all, for the Trump pre-election campaign on the Internet, and the campaign in support Brexit it is the same company that investigates Big Data: Cambridge Analytica and its director Alexander Nix.Anyone who wants to understand the nature of the voting (and the nature of what awaits Europe in the coming months), must begin with the remarkable events that occurred in 2014 in the British University of Cambridge. Namely, at the Kosinski Department of psychometrics.

A look into the history

Psychometry, which is sometimes called psychographics, is an attempt to measure the human personality. In modern psychology, the standard is the so-called "ocean method." In the 1980s, two psychologists have proven that each trait can be measured using five measurements. This so-called "Big Five": openness (how much you are ready for the new), integrity (how much you are a perfectionist), extraversion (how you feel about the society), benevolence (as far as you are friendly and willing to cooperate) and neuroticism (how easy it is to get you from itself). By these measurements, you can understand exactly what kind of person with dealing in what its desires and fears, finally, how he can lead. The quality assurance software problem was data collection: to understand something about the man, and he was required to fill out an extensive questionnaire.But then came the Internet, and then Facebook, then Kosinski.In 2008, a student from Warsaw Michal Kozinski started a new life: he entered the prestigious English Cambridge Center psychometrics, Cavendish Laboratory, the very first laboratory in the world of psychometrics. With his classmates, he invented and launched a Facebook application called MyPersonality. The user is asked to respond to a long list of questions ( "Do you easily deduce from itself in a state of stress? Do you have a tendency to criticize others?"), Then received his "personality profile, " and the creator of the application received invaluable personal information. Instead of the expected data on a dozen classmates creators received information on hundreds, thousands, and then millions of people. Two doctoral collected thus the largest "harvest" of the data in the history of psychological research and mobile software development.

The process that Kosinski associates developed over the next few years is quite simple. Firstly, it obtains a list of test questions online test. From the responses to it, scientists calculate the personal values of the test. Further, Kosinski with the team studied the effects of the test: likes and reposts to Facebook, as well as gender, age, and place of residence. So researchers are due. A simple analysis of the data on the network can get unusual conclusions. For example, if a man is subscribed to a MAC cosmetics brand page, it is a high probability that he is gay. On the contrary, a strong indicator of heterosexuality - if a person Like hip-hop group Wu-Tang Clan in New York. Fan Lady Gaga is very likely an extrovert, and if the person is putting "like" to philosophical post - introvert.

The analysis of 68 likes on Facebook is enough to determine the person's skin color (95% probability), his homosexuality (88% probability) and the commitment of the Democratic or Republican Party.

Kosinski and his colleagues constantly perfected its model. In 2012, Kosinski proved that the analysis of 68 likes on Facebook is enough to determine the person's skin color (95% probability), his homosexuality (88% probability) and the commitment of the Democratic or the Republican Party of the USA (85% probability). But the process goes further: intellectual development, religious preferences, addiction to alcohol, smoking or drugs. These even allowed to know whether the person's parents divorced before his age or not. The model has been so good that it became possible to predict the person's answers to certain questions. Drunk with success, Kosinski went on: model could soon get to know the person better than his colleagues after ten studied likes.  After 70 likes - better than his friend. After 150 likes - better than the parents.After 300 likes - better than a partner. With an even greater amount of learning activities can be found on the person better than himself. On the day when Kosinski published an article about his model, he received two calls: the complaint and a job offer. Both calls were from Facebook.

Best regards, 
Your Ardas Team