Using diseases and businesses as metaphors to understand gangs create interesting insights. According to new research from Yale and the West Point Military Academy, graph theory can be applied by law enforcement agencies. Armed with graphs, they can identify at risk individuals or target gang leaders.

Cristopher Wilderman and Andrew Papachristos from Yale argue in a recent paper that in poor neighborhood, the most important factor to explain the probability of being killed is not race, gang membership or physical proximity to a murder victim. It is simply about who you know. Their study shows that one’s social network influences the probability of being killed. Among the network characteristics they considered are how people are connected, the structure of the overall network, the types of behaviors occurring in the network and an individual’s position in the overall structure.

This means that police forces could potentially use graph theory to identify and protect at risk individuals.

One of the most common application of graph theory in enterprises is Organizational network analysis. We all know organizational charts, these nice pictures that describe the hierarchical relationships within people. These charts do not account for the way information flows within the organization and people collaborate. Networks offer a much better picture of real world businesses. Companies like Orgnet or Activate Networks use Organizational network analysis to identify :

• influencers : people who can get things done by leveraging a network of people throughout the company;
• bottlenecks : people who are hurting the overall flow of information in a company;
• areas where collaboration can be improved : people who are under-connected or lack important strategic connections to accomplish theirs goals;

A team at the West Point Military Academy is using the same logic to help law enforcement agencies understand gangs. After all, criminal organizations are organizations that share many traits with businesses : they have leadership, command structures and common objectives. The problem is that they seldom publish an official organizational chart. That is where Orca (Organization, Relationship, and Contact Analyzer) comes in : this software can use data acquired from arrests to deduce the network structure of a gang.

This in turn can be very useful. Armed with the knowledge of the gang structure, police forces can target influential individuals. People they should collect information on to learn about the whole organization or people they should arrest to disturb the whole network.

Even if you’re not fighting gangs (or running one), you can apply graph technologies to better understand your data. What do you want to use graphs for?