Photo: Mark Greenwald, chief of research and planning, Florida Department of Juvenile Justice
In Florida, more than 85,000 kids flood the state's juvenile justice system every year. But this year, officials hope to shrink that number using new technology that examines key historical data to forecast the future.
Announced Wednesday, April 14, the Florida Department of Juvenile Justice has implemented IBM SPSS predictive analytics software to analyze predictors such as past offense history, demographics, gang affiliation, peer association and home life environment to find out which kids have a higher chance of recidivism. The idea, then, is to target high-risk cases with programs that address specific needs for rehabilitation so they don't stumble back into the system.
"My job is to put myself out of business," said Mark Greenwald, chief of research and planning at the department. "It's important that we direct what resources we have to kids who need us the most and use the software to build upon the decision-making process."
In years past, to predict delinquency numbers the department used Excel, which had limited functionality that allowed the department to look "in the rearview mirror," Greenwald said, but wasn't advanced enough to help state officials see clearly ahead.
The new IBM SPSS analytics software will allow them to better intervene in juveniles' lives earlier, and match up those youths with programs based on their specific situations, such as drug abuse, mental health issues or criminal activity.
A Calculated Guess
Employed in various sectors including health care, retail and public safety, predictive analytics uses a wide range of techniques -- such as data mining and statistics -- to predict future events or behaviors. In financial services, for example, the concept of credit scoring represents one of the most popular applications: Lenders examine credit histories and other financial data to help determine whether a customer would make credit payments on time.
Proponents say predictive analytics can turn gut hunches into hard numbers. In the private sector, such data may be used to price products, maintain inventory and boost profits. In the public sector, the data could improve efficiency and services by targeting problem areas and predicting probable outcomes.
For instance, a public school system in the U.K. used an IBM predictive analytics solution to identify and support students who could be subjected to violence, blackmailing and bullying, according to Erick Brethenoux, a predictive analytics expert at IBM. In one police department, he said, the software helped officers make connections between spikes in crime at the end of the month when payments are due, coinciding with phases of the moon. Last month, the Ministry of Justice, one of the U.K.'s largest government departments, selected IBM's predictive analytics technology to analyze 4 million prisoner records to predict the likelihood that they will commit crimes upon their release.
In the past few decades, TV shows like CSI and Numb3rs have given predictive analytics more mainstream appeal, Brethenoux said. "More people believe that mathematics can really help in solving problems," he said. "They accept the fact that you can solve crimes in labs."
Most of the time, organizations already have the data, but the software helps employees make sense of it, and use it to influence decisions. Once that happens, Brethenoux said, agencies and organizations can look forward to improved results as well as return on investments.
As a top long-term analytical priority, more than two-thirds of senior managers in the U.S. and the U.K. want to predict behavior, actions and decisions and use the data to make quick choices, according to a recent Accenture survey. But the survey of 600 senior managers at more than 500 blue-chip organizations also found that more than