Anyone who has worked with mobile-phone data knows how incredibly useful such information can be, even when it’s anonymous. It is amazing—but at the same time frightening—what massive quantities of spatio-temporal data points from mobile phones can tell us about ourselves, our lives, and our society in general.
Mobile phones know where we are and when, and whom we talk to. In some cases they even know when and in what amounts we add credit to prepaid phones, which in some places is a good proxy for how much money people have. All this data can be harnessed for the public good (see “Big Data from Cheap Phones”). In countries where even population estimates are hard to get, mobile phones constitute a unique source of information.
Recently, the telecom operator Orange challenged researchers around the world to analyze “anonymized” mobile-phone data sets from Ivory Coast and see how the information might be used. The data sets are based on more than two billion records of communications between five million customers in the African country.
This “data for development” challenge—the first of its kind—has been received with tremendous enthusiasm. Over the last six months, hundreds of researchers have proposed ideas that are creative, original, and useful. Among many others, they suggest ways to respond to emergencies, improve health, optimize transportation infrastructures, monitor development policies, prevent violence, and anticipate the spread of diseases such as meningitis, malaria, or cholera.
Unfortunately, without proper care these kinds of data sets can be misused, and making the information available could compromise people’s privacy. A few data points suffice to identify most customers, even if their names are stripped from records. But at the same time, those data points may save their lives, or at least help make those lives better and safer. These trade-offs should be worked out and debated so that we can benefit from data in a way that respects the interests of all.