The term “big data” has caught on in recent years as a way to talk about massive quantities of calculations, figures, and computations made by extremely high-tech machines like the supercomputers at Oak Ridge National Laboratory.
Now a national nonprofit group that includes a professor from UT’s Tickle College of Engineering has found a new use for such data: crime solving.
“The idea is to try to help solve cold cases through the use of big data and computer analytics,” said David Icove, UT’s UL Professor of Practice in the Department of Electrical Engineering and Computer Science. “Data can reveal what the human component might not be able to distinguish.”
The Murder Accountability Project was formed as a way to use massive amounts of data to find patterns and predict possible outcomes for cases that might otherwise be stumping law enforcement.
The group, made up of retired FBI criminal profilers, journalists, criminologists, and academic researchers, has developed an algorithm to detect serial murder cases, based on FBI Uniform Crime Reporting data on murders.
Icove, who serves on the group’s board of directors, is a former FBI instructor. He wrote the definitive book on fire investigations and on various areas of criminal forensics.
In fact, when he left the FBI, the agency was forced to suspend the course he had taught because no replacement could be found who matched his expertise.
Locally, he worked on the investigation into fires that swept through the McClung Warehouses in 2007, helping determine the cause and timeline of the blaze.
On the accountability project, Icove and the team developed an algorithm that could be applied to known data to help determine patterns.
That concept was recently put to use in Cleveland, Ohio, where the data was used to map out 60 cold cases involving the murder of women over the past 12 years.
Through the use of their algorithm, the group was able to identify several clusters of cases that were marked as likely being the work of the same killer.
“It really is a perfect example of how big data analytics can assist both law enforcement and the criminal justice research community,” said Icove.
David Goddard (865-974-0683, firstname.lastname@example.org)