Cell phone data analysis dials in crime networks
By Rachel Ehrenberg
Sometimes not picking up the phone can be as incriminating as spilling the beans on a wiretapped call. After a recent string of robberies in Italy, a new forensic tool that makes it easy to explore reams of cell phone data revealed an incriminating pattern: Leading up to and after each robbery, there were flurries of calls between members of a gang suspected in a series of gun thefts, car thefts and supermarket stickups. But in the end, it was silence that betrayed the gang: There were no calls during the minutes when the crimes were being committed.
The new data analysis tool called LogAnalysis makes it especially easy to visualize the relationships among conspiring suspects as revealed by their phone calls. Developed by scientists at the University of Messina in Italy, the program takes algorithms that researchers typically use to investigate relationships among organisms in an ecosystem or the flow of information in a friendship network, and brings that math together in a manner tweaked specifically for investigating crime. Described in a paper posted March 7 at arXiv.org, LogAnalysis is being tested with some criminal cases in Italy, although the researchers will divulge little about the real world cases.
To begin, investigators import all the call data into the program, which transforms the information into a diagram of people connected by phone calls. There are features that make it easy to see who called whom the most and to identify clans and go-betweens. Investigators can also trim people and calls out of the visualized network when deemed irrelevant to a case.
But it was a temporal analysis feature that helped alert investigators to the supermarket holdup crew’s nefarious doings, says police detective and study coauthor Salvatore Catanese. The gang was in rapid communication leading