“It was ‘Shorty’ that shot, stabbed, and ran over me,” the victim told police. That was all he knew, but that was enough for CopLink, software used by the Tucson police to help them solve crimes. Having received this fragment of information at 10:00 A.M., they were able to have a suspect in custody by 5:00 that evening.

Described as a “Google for cops,” CopLink, which costs about $60,000, searches through 911 records, arrest records, and police reports for names, license plate numbers, lists of stolen items, fingerprint records, and so on, to find links, suspects, and even leads.

For example, with an approximate height, weight, hair color and an “N” in the a license plate, police can quickly find possible suspects in an arson case. In the past, law enforcement had to search through each database separately and could find links only after many hours of studying all clues from different angles. Just going through paper-based fingerprint records can be a daunting task. Boston, for example, has 350,000 sets of fingerprints on file. It takes a long time to find anything in that many sets of paper records, but it takes only seconds with a computer.

CopLink employs neural network techniques to associate information and find patterns. The program also creates a detailed audit trail. There are two good reasons for having such a capability. First, officers are able to demonstrate the facts that led to the questioning or arrest of a suspect. Second, it allows the law enforcement agency to monitor the system for signs of abuse and to stop it quickly.

CopLink has six components: one to associate relationships, a second to search through all the information, a third to represent its findings graphically, a fourth to manage security, a fifth to allow the use of updated information and to share information with other agencies, and a sixth that allows officers to wirelessly connect to the system when in the field.