At the recent 911 Critical Issues Forum—presented by the National Emergency Number Association (NENA)—Mike Reade, public safety specialist for IBM’s Integrated Smarter Solutions Team, spoke about the need for public-safety agencies and public-safety answering points (PSAPs) to embrace data analytics.

“There is so much more information coming at us today than just a few years ago,” Reade said. “You can’t possibly grasp all of it [without analytics]. The traditional way of doing things doesn’t cut it anymore.”

Harnessing data will provide agencies with a better understanding of why an event occurred, which will help them make better decisions should that type of event happen again, Reade said. More importantly, the information could be used to help agencies predict and even prevent incidents from occurring.

“To make a prediction, you really need to be able to incorporate a wide variety of variables. These often are pieces that don’t exist in our traditional systems,” Reade said. “It’s not enough just to look at crime-history records.

“You also need to incorporate data into your predictive-crime model that helps you understand what else was happening when these crimes took place. What was the weather like? Was there a public event going on? Was it the day when a lot of people received their government checks? Predictive analytics can help identify the trigger factors.”

Reade offered the following example. Let’s say that police department discovers a spike in crime in a certain sector of the city by reviewing its crime-history data. This is good information, because it lets the department concentrate its resources where they are needed most. But, if the department had been able to overlay data from other sources, it might have discovered that the crimes were occurring in an area that simultaneously generated numerous 311 calls about street-lamp outages.

“In that case, what is the better course of action—putting more cops on the street or fixing the street lamps?” Reade asked. “Data analytics helps you connect the dots between numerous databases and helps you become aware of non-obvious relationships.”

Another example concerns unsolved crimes, according to Reade. Data analytics will help agencies detect patterns and common denominators that will help them solve more crimes, and they can be used to predict both repeat offenders and repeat victims, he said.

“Understanding an offender, understanding what patterns an offender fits into and the history of those patterns, and what factors have contributed to those patterns—based on this analysis, when an offended is realized from corrections into the real world, you’ll be able to predict the likely behavior that they will exhibit,” Reade said.

Reade pointed to Miami-Dade County as an example of how analyzing repeat offenders can pay off. The county had numerous unsolved property crimes in its case files. The first step was to create a “huge matrix” that identified numerous characteristics for thousands of crimes—for example, where they occurred, when they occurred, and what was damaged or stolen. Then, the county cross-matched data it had on known offenders.

“That allowed them to do a scoring of the crimes and the offenders, to identify a match,” Reade said. “Then they started knocking on doors. By the time they got to number five on the list, they made an arrest. Ultimately, they closed 25 cases that they never would have been able to solve otherwise. The key is in determining whether a crime fits a pattern.”