Art of Analytics: Corona of Calls

Learn Data Science
Moderator

corona-of-calls.png

 

The Insights

 

What would you see in a giant corona, or a large circle of light, with colorful leaves of varying lengths sprouting around it? This corona depicts a network of calls made in a geographic area prior to a terrorist incident.

 

Frantic calls are usually made after an incident occurs by people affected by it. However, calls made before the incident are also informative and can contain clues about the perpetrators. These are the patterns of calls that law enforcement and intelligence agencies want to identify, isolate, and use to locate the perpetrators as well as to prevent future incidents from occurring.

 

Each leaf shows a distinct network of calls that were made. Each node within a network, or leaf, is a call that was made to others within that network. Certain calls spanned networks, and some nodes appeared to have made calls of great frequency before the incident and then fell completely silent in its aftermath.

 

This graph of connections can also be used to identify unknown individuals who were called from known individuals' numbers. These known numbers and their connections can be isolated within geographic regions to determine the likely pool of suspects, which can then be used for further investigations.

 

The Analytics

 

This visualization was created using Teradata Aster Analytics. We used detailed cell phone call data (hundreds of gigabytes or terabytes, typically) that was loaded into Aster Analytics for analysis. Aster's native graph algorithms (e.g., Closeness) were used to determine the structure of these calling networks and the strength of the associations within these networks. A visualization of the underlying node-edge table that depicts the network linkages was then rendered using native visualization functions.

 

The Benefits

 

At least three benefits are realized through this advanced analytics implementation:

 

First, using calling connections to isolate unusual calling patterns ensures that when such patterns are repeated, there is a higher likelihood of thwarting harmful incidents before they occur.

 

Second, identifying a nefarious circle of connections ensures that the likely perpetrators are quickly identified, helping to bring the culpable parties to justice.

 

Third, call records of the suspected perpetrators can be identified and parsed prior to an incident to determine the nature of specific security threats or code words that are likely to be used in terrorist operations.