A paper (also available here) published yesterday in Nature analyzes moving patterns of humans based on position data provided by a European mobile network operator. While the paper itself is very interesting and provides new insights it also raises serious privacy concerns, and maybe even legal troubles for the telco in question.
Although the paper claims the position data has been anonymized, strictly speaking this has not been the case. For a series of events involving a set of individuals, anonymity does not only mean that for any event it is not possible to identify the individual, but also that it is impossible to establish correlations like "the person associated to event A also caused event B". If such correlations can be derived from a data set one rather speaks of pseudonymity. Of course, to be able to tracking moving patterns of individuals, the authors had to have access to them.
This may seem to sound like nit-picking, but has some serious implications. Let me explain it with a short example. Suppose you lived in a small town that entirely lies within a single mobile network cell. If you have access to pseudonymous position data and know that your unfriendly neighbor spent his summer vacation on some lonely island a few hundred miles away, you can immediately reveal the rest of his movements. While this is a drastic example, people have been able to identify individuals from much weaker correlations.
While it is hardly the fault of the authors of the paper, the telco providing the data (which is not revealed, but you can make some guesses based on subscriber counts and the movement patterns shown in Fig 1a) should be asked some questions. And, if said telco operates in a country like Germany, where proper anonymization is mandatory for subsequent scientific use, some legal troubles might loom as well. Especially, since we are currently being rocked by another telco privacy scandal.
Reference: M. C. González, C. A. Hidalgo, A.-L. Barabási. Understanding individual human mobility patterns. Nature 453, 779 (2008).
Copyright 2006--2011 Hendrik Weimer. This document is available under the terms of the GNU Free Documentation License. See the licensing terms for further details.