PredPol Predicts Crime

By Christie Vaughn

Putting a cop at the crime scene before it happens with a mathematical formula sounds like it came straight out of a Hollywood movie, but with new technology, it's now possible.

George Mohler, an assistant professor in the mathematics and computer science department at Santa Clara University, is one of the masterminds behind Predictive Policing, a new computer software. His company, PredPol, may be the next era of crime prevention.

Since police departments nationwide are facing budget cuts, this new resource allows them to be more effective and responds to public demand for crime prevention without spending more money.

Mohler started working with a team of researchers from the University of California Mathematical and Simulation Modeling of Crime project, funded by the U.S. National Science Foundation. The UC MaSC project has one goal: to develop a statistical model for crime.

With the help of research from this group, Mohler was able to come up with a mathematical equation to predict crime based on the same formula currently used to predict aftershocks for earthquakes.

After publishing papers about his theory, Zac Friend, Santa Cruz Police crime analyst, reached out to Mohler about implementing the algorithm. The predictive policing program was put to the test.

"We decided to try the predictive policing program because our department had experienced about a 20 percent decline in staffing coupled with a 30 percent increase in call-for-services," said Friend.

Since implementing the program, there has been a decrease in Santa Cruz crime, including residential, commercial and vehicular burglaries, dropping all burglaries by 27 percent.

Another experiment with Los Angles Police Department saw similar drops in crime.

Mohler presented the statistical model to one of his classes at Santa Clara. One of his calculus students, Omar Qazi, approached Mohler after the discussion.

"I knew I could make his software better," said Qazi.

Qazi and Mohler started working together to further develop the software.

"(Qazi) has ultimately made this software better, and I would work with him regardless of his age because he is a unique, talented individual," said Mohler.

PredPol analyzes the times, dates and places of recent crimes. It updates constantly when new crimes occur and then finds patterns that forecast high-risk times and locations for future crime. The cost to implement PredPol depends on the population of the city, but it is always less than the cost of hiring an officer for a year.

PredPol may violate the Fourth Amendment, however, which protects against unreasonable searches and seizures.

Professor Andrew Ferguson from the University of Pennsylvania School of Law wrote an article called "Predictive Policing and Reasonable Suspicion," examining predictive policing and the possibilities of how it will impact reasonable suspicion.
"In the future, predictive policing will affect the Fourth Amendment reasonable suspicion analysis," said Ferguson. "How it affects it and whether these changes weaken or strengthen Fourth Amendment protections remains unclear."

Friend said, "The program is not a replacement for an officer's talents, intuition or experience. It is simply a tool that helps them do their job."

The Sponsored Projects Office at the university assists in proposal, funding, developing project ideas and submitting completed project plans. Mary-Ellen Fortini, director of Sponsored Projects, reviewed the grant Mohler submitted for PredPol.
"We have to balance between the gain and the risk - do the risks outweigh the benefits?" said Fortini. "In this case, predictive policing benefits outweigh the risk."

Major cities around the United States including San Francisco, Palo Alto, Mountain View, Los Angeles, Santa Cruz, Los Gatos, and as far away as Kent, England are using PredPol today. The future of the war against crime now blinks on a computer screen.

Contact Christie Vaughn at cvaughn@scu.edu.