On Dec. 13, 2003, the 1st Brigade Combat Team of the U.S. Army’s 4th Infantry Division discovered Saddam Hussein hiding in a narrow, dark hole under a two-room mud shack on a sheep farm in the former Iraqi dictator’s hometown of Tikrit.
Guiding the military’s months-long manhunt was a complex visualization of tribal and family members with close ties to Saddam. By applying the principles of social science to analyze these networks, it became possible to uncover specific individuals and relationships that ultimately revealed Saddam’s hiding place.
Now, imagine trying to uncover these relationships in a mountain of data on a terrorist network with 20,000 members, or “nodes,” and more than 100,000 relationships. A number of new products are doing just that by sifting through clutter to reveal key connections within terrorist groups and possibly to predict their behavior. “No longer can analysts use just an organizational chart to describe an enemy’s configuration,” Lt. Col Brian Reed wrote in a recent issue of Parameters, published by the U.S. Army War College. Reed is one the architects of the plan that led to Saddam’s capture.
At the heart of this boom is a discipline known as social network analysis (SNA), which uses mathematical algorithms to illuminate the patterns of the ties that link people, groups and organizations. By understanding these relationships, analysts can determine the most central figures in the network, their relationships and other important factors. Although this methodology has been with us for about 50 years, it began to garner fresh interest with the advent of supercomputers in the 1970s. The modern desktop computer is taking SNA to a new level.
Sentinel Visualizer, developed by Vienna, Va.-based FMS Advanced Systems Group, is one of several new programs that aim to bridge the gap between merely displaying networks and fully analyzing them. It is able to find the shortest path between two entities while filtering data regarding specific people and places, relationships and the status of individuals who may be incarcerated or dead.
Other algorithms included in the program can measure the relative importance of a node in the network, factoring in time and location, among a variety of considerations.
FMS President Lake Chung said the program also provides users with the ability to augment other types of visualization software and layer new data as it is fed into the system. The system downloads easily, he said. and can be customized for individual, group or enterprise needs. Interest in this capability is growing, Chung said.
Until now, customers interested In FMS’ suite of SNA tools have come mainly from the defense and intelligence communities. One recent investor was In-Q-Tel, the CIA’s technology-focused venture capital arm. Chung said Sentinel Visualizer enabled the CIA to move away from the traditional mode of analysts looking at information in "rows and columns". FMS is now marketing Sentinel Visualizer to civilian law enforcement agencies and large corporations, including those that manufacture pharmaceuticals.
The Pentagon is using SNA software to analyze a wide range of networks including some within its own walls. In 2005. the former Office of Force Transformation, working with the Defense Acquisition University, used social network analysis in an effort to grasp how decisions are made throughout the Pentagon.
Similarly, the U.S. Coast Guard is seeking bids for a social network analysis tool to help the organization better understand and improve communication between its various commands and outside agencies.
These projects, which are based mainly on raw data available through open sources, don’t face the same hurdles as SNA efforts focused on enemy networks. When analyzing terrorist networks, interpretations must be built on potentially unreliable information, or with key elements fragmented or missing.
Work is moving forward, however, and SNA tools and techniques are earning new converts daily. For example, the Defense Threat Reduction Agency has asked social science researchers to develop an SUA framework and methodology specifically geared toward networks involved with nuclear, biological and chemical weapons.
Stephen Borgatti, a consultant on SNA, said the software can create two- and three-dimensional charts, temporal displays and geographic depictions that quickly reveal the kinds of relationships formed in these highly technical areas of study.
"Visualizations like these can use different colors, sizes and shapes to denote important members of a network, the strength of the relationships between the members, and also which players tend to be the most isolated.” he said.
Another program trying to bridge the analysis-use gap is Network Analyst, which is being developed by Blue Spider Analytics of King George, VA. This software, slated to be released this year, will attempt to bring together algorithms that deal with financial, telecommunications and other networks into a library of algorithms analysts can use.
Michael Pastore, Blue Spider’s principal engineer, said the program will allow users to add new SNA or layout algorithms as they are developed. With deeper analysis, visualization capabilities will also adapt to meet the display needs, he said.
Jennifer Johnson, an assistant professor of sociology at Virginia Commonwealth University. said visualization makes SNA products less cumbersome to use than earlier-generation systems that either display data with little analysis or are grounded in academia.
“There are a lot of commercial tools out there that do visualization. They can put up pictures, but a lot of them are not really doing analysis, they’re just doing display,” she said. "And what the commercial products suffer from a little bit is that they are mostly cobbling together, for example, pieces of UCINET or Excel, rather than starting from scratch to be able to do a truly Web-based or windows-based SNA tool.”
Another issue, Borgatti said, Is that many SNA tools are built for reasons other than defense and intelligence, so analysts in these fields must modify them with additional algorithms to obtain the desired result.
As such, users should refrain from impulsive action without input from analysts or subject matter expert;, Johnson said.
"One common misperception about SNA is that the picture will somehow reveal all," Borgatti said, adding that the software is not a cure all when it comes to revealing social networks.
"We've passed the awakening phase, and now we’re in the buzzword phase,’ Blue Spider’s Pastore said. "Within the defense community there is limited understanding of the potential of SNA, and many of the tools on the market provide only a limited capability. Thus, they can erroneously shape the perceptions defense officials have on the methodology and its power."
Although SNA remains a hot topic at the Defense Department, Borgatti cautions that stakeholder expectations must be managed carefully.
"Not only does SNA involve understanding networks, it’s about using that knowledge to predict how they will evolve. We have to ensure that misconceptions about SNA tools and methodologies don’t lead to a backlash against them."