Constantly collected via mobile phone apps, fitness trackers, credit card logs, websites visited, and other means, is data about our habits and movements. It seems that data collected from acquaintances and even strangers can predict your location, even with it is turned off.
An associate professor of physics, mathematics, and computer science at the University of Rochester, Gourab Ghoshal says that “switching off your location data is not going to entirely help.”
As a result, Ghoshal and colleagues decided to test just how far-reaching a person’s data might be. So, they applied techniques from information theory and network science to find out. Most importantly through their research they discovered that switching off your data tracking didn’t really help. In other words, an individual’s mobility patterns could still be predicted with surprising accuracy based on data collected from their acquaintances, even if they had turned off their individual data tracking and didn’t share their own information.
Meanwhile, “almost as much latent information can be extracted from perfect strangers that the individual tends to co-locate with,” says says Ghoshal, which is even “worse.”
FRIENDS AND STRANGERS
Ghosal and his research team investigated four datasets:
- three location-based social network datasets composed of millions of check-ins on apps such as Brightkite, Facebook, and Foursquare
- one call-data record containing more than 22 million calls by nearly 36,000 anonymous users.
In addition, they developed a “co-location” network which enabled them to distinguish between the mobility patterns of two sets of people:
- those who are socially tied to a person, such as family members, friends, or co-workers.
- A stranger tied to an individual, who are at a location at a similar time as the individual. For example, you work in the same building but are at a different company, or parents whose children go to the same school, or you shop at the same store.
Most importantly, the researchers discovered that movement patterns could be predicted using social ties to an individual with 95% predictability. They did this by applying information theory and measures of entropy—the degree of randomness or structure in a sequence of location visits
But surprising they were able to achieve an 85% predictability of an individual’s movement using stranger ties
DATA TRACKING’S SLIPPERY SLOPE
Data tracking and the ability to predict where an individual or group is at any given time does have its benefits. For instance, during a pandemic where controlling the disease and stopping the spread is able via contact tracing based on mobility patterns. In addition, there is urban planning would also benefit to taking a broad look at groups and individual movements to see where money needs to be spent. Lastly, data mining to offer tailored recommendations for restaurants, TV shows, and advertisements for many consumers is appreciated rather than getting irrelevant information.
On the other hand, Ghoshal warns “that data mining is a slippery slope.”
Firstly, because as the research indicates, individuals maybe unwittingly providing information about others by sharing data via mobile apps.
Secondly, it’s a slippery slope because “people should be aware of how far-reaching their data can be, namely with implications for surveillance and privacy issues and the potential rise of authoritarian impulses,” Ghoshal continues.
The answer isn’t as easy as just switching off your phone or going of grid. But it will mean that more talks about how people are collecting your data and using is required. And laws and guidelines set in place for protection.
