Nowhere does the power of Big Data cause more concern than when in the hands of the National Security Agency. Though it's not the idea of Big Data that disturbs people, so much as it is the idea of mass surveillance. Privacy panic is all the rage, and while people fear terrorist attacks, they don’t necessarily believe preventing one is worth someone reading their e-mails.
The case for using the Big Data gained from mass surveillance only holds up if it actually works though. In a Financial Times blog, Zeynep Tufekci argues that analytics are, by their very nature, the wrong tool to be using. Big Data analytics, she says ‘when conducted on massive datasets can be powerful in analyzing and identifying broad patterns, or events that occur regularly and frequently, but are singularly unsuited to finding unpredictable, erratic, and rare needles in huge haystacks.’
Tufecki argues that government forces would be better prepared to prevent attacks by paying attention instead to the causal chain that leads terrorists on their path. Which, to continue her metaphor, isn’t a very good way of finding a needle in a haystack either. The best way to find a needle in a haystack is with a giant magnet, though it’s unlikely that this would be a successful way of catching terrorists.
This is besides the point anyway, as the two things are not mutually exclusive. One of the advantages of Big Data is that it can be used to build profiles of potential terrorists and locate areas with large numbers of people who match these. One of the main tools being used against ISIS is social media analytics. ISIS is renowned for their social media prowess, with the use of Twitter and Facebook central components of a recruitment drive that has seen them become the most popular organization for fighters coming from foreign countries.
The terrorist group and its supporters sends an estimated 200,000 tweets a day, a ripe number for analysis that can be used to leverage a number of actionable insights. Big Data taken from social media can establish what motivates terrorists and determine the characteristics of potential recruits. A massive data mining project conducted by the Qatar Computing Research Institute in Doha analyzed data from social media to find the origins of support for ISIS, looking at over 3 million tweets sent over a three month period, from which they created an algorithm that could identify user sentiment to an 87% level of accuracy.
The article is also flawed in that Tufecki makes a number of assumptions as to the success, or lack there-of, of Big Data as an anti-terrorism tool. She cites several examples of terrorist atrocities in the West as examples of Big Data’s failings, but neglects to mention times it may have been successful, mainly because she doesn’t know.
The simple truth is that there is no easy solution for terrorism, and lone wolves are always going to occur as they are incredibly difficult to prevent. However, it would be foolish to disregard Big Data as a tool in the fight.