Complexity and Social Networks Blog

« December 22, 2005 | Main | January 4, 2006 »

30 December 2005

Social network analysis, the NSA, and “pattern analysis�

The story about the NSA eavesdropping program has received a lot of attention over the last week. The follow up story has received somewhat less attention, but may be more important, see story from December 24 NYT: Spy Agency Mined Vast Data Trove, Officials Report (by ERIC LICHTBLAU and JAMES RISEN):

“What has not been publicly acknowledged is that N.S.A. technicians, besides actually eavesdropping on specific conversations, have combed through large volumes of phone and Internet traffic in search of patterns that might point to terrorism suspects. Some officials describe the program as a large data-mining operation.�

"All that data is mined with the cooperation of the government and shared with them, and since 9/11, there's been much more active involvement in that area," said the former manager…

This is a remarkable story, and raises some interesting questions: (1) exactly what data are telecomm companies sharing with the government; (2) what could usefully be gleaned from these data; and (3) what are the privacy implications?

There is a lot more we don’t know about this story than we do know, but it is worth beginning a discussion on the value and the costs of these data under different scenarios of exactly what information is being shared. My next few entries will aim to begin a discussion on these issues, grounded primarily in a social network perspective.

Briefly, what data do telecomm companies have? Focusing on the telephone data, for now, they have (1) phone log data; (2) varying amounts of locational information for cell phones; and (3) varying amounts of information linking individuals to particular phone numbers (e.g., not so much for some pay as you go phones, more for other types of phones). My understanding is that little remains of the bits that flow over the network (i.e., the content).

These are thus a type of social network data, along the lines of my preceding entry on the “behavioral flows� of relationships. That is, for any given dyad one can observe the timing and duration of calls.

Whose phone data is being tracked? It is not clear from the article. Clearly, the focus is on international communication (domestic to international, and international to international calls routed through switches that are on US soil). Is purely domestic communication also being tracked? The article suggests not:

“This so-called "pattern analysis" on calls within the United States would, in many circumstances, require a court warrant if the government wanted to trace who calls whom.�

This sentence is ambiguous, however—e.g., given that the sharing of data by the telecomm companies is voluntary, what are the statutory limits on their sharing data with the government? Is there a prohibition on a telecomm company voluntarily handing over information to the government regarding one of their customers’ phone logs? I do not know of such a prohibition, but if a reader does, please do comment.

For the next entry: Given that these are essentially social network data, from what we know from the research on social networks, what insights might they yield?

Posted by David Lazer at 11:19 PM