📜 ⬆️ ⬇️

The cardholder can be distinguished from a million list of all four purchases

image

Researchers from MIT exposed several myths about privacy in modern society. They found that even incomplete data on four purchases made using bank cards, allow us to identify a person in 90% of cases. You need to know only the date and place of purchase. These data were compared with a list of transactions for three months, committed by more than a million people.

Moreover, if we use the rough data on the prices of purchases (the cost of the purchase with an accuracy of 50%), then the number of reference points is reduced to three, and the accuracy rises to 94%. That is, for example, having armed with one check thrown away by you, one photo of your breakfast in a cafe from Instagram and one of your tweets with a message about buying a new gadget, you can almost certainly identify you in the list of transactions among a million other buyers.

For the study, a list of transactions was used, including the name and location of the stores, as well as the date and amount of the purchase. Purchases made from the same bank card were marked with the same id. While working on the project, the researchers tried a different number of reference points, from two to five. Even when using two points without information on the purchase price, identification of a person becomes possible in 40% of cases. Using five points leads to results close to 100%.
')
The question arose - how could the data set be modified so that it was still of interest from the point of view of economists and marketers, but at the same time improving the privacy of customers. To do this, the researchers tried to make information less accurate, by enlarging the time interval for buying up to a week, and summarizing up to 150 stores in a certain area. And still, it was possible to single out a person from the total set of transactions with a probability of 70% for the four reference points where the purchase price is indicated with an accuracy of 50%.

Modern means of processing big data from such seemingly impersonal information as the base of purchases, or the base of movements and use of mobile phones , make unexpectedly accurate conclusions. Perhaps ordinary citizens should not be afraid of such opportunities. But you do not need to harbor any illusions about your anonymity, if for some reason this is important to you. Especially when you actively use social networks, mobile phone and bank cards.

Source: https://habr.com/ru/post/365971/


All Articles