Predictive Policing Symposium: Technical Breakout Session
John Morgan, Ph.D.,
Office Director, Office of Science and Technology, National Institute of Justice
When NIJ and BJA convened the Predictive Policing Symposiums, seven police agencies were developing strategic plans for doing predictive policing in their jurisdictions.
First Predictive Policing Symposium
At the First Predictive Policing Symposium, Dr. John Morgan emphasized that the technical breakout session would provide an opportunity to discuss what information should be applied to these programs.
Morgan offered the following definition of predictive policing from the Harvard Executive Session paper: “Predictive policing refers to any policing strategy or tactic that develops and uses information and advanced analysis to inform forward-thinking crime prevention.”
The group commented on specific elements of the definition:
- “Advanced analysis”: Some felt the word “advanced” was not appropriate because simple methods could be used for prediction. Another felt that “analysis” is a powerful word that can stand on its own.
- “Crime prevention”: Some suggested that the definition should include any activity in which police might engage, not just crime.
- “Forward thinking”: Some questioned if this needs to be included. Others noted that “forward thinking” implies backward thinking.
- “Predictive”: A number of participants felt that “predictive policing” might not be the most appropriate term. Suggestions of terms to consider included “proactive policing,” “preventative policing,” “adaptive policing,” “evidence-based policing” and “data-driven policing.”
Morgan reintroduced the five elements of predictive policing presented in the opening panel session (The Future of Prediction in Criminal Justice: Defining Terms and Introducing Issues), and participants offered their comments:
- Integrated information and operations: There are no silos to effectively integrate information and operations. One participant noted that strong leadership will be necessary to develop network-centric policing. Other participants said that there are analysts producing good products, but their products are not used because of their civilian status.
- Seeing the big picture: Prevention is as important as response, and every incident is an information-gathering opportunity. Someone noted that with definitions, there is a top-down focus; to welcome a network-centric environment, a bottom-up approach is necessary.
- Cutting-edge analysis and technology: There is a wealth of tools and technology already available and it is imperative that departments learn how to use them.
- Linkage to performance: Most participants agreed about the importance of this relationship.
- Adaptability to changing conditions: This concept highlights the need for flat-networked organizations, training in how to adapt to strategies based on information and high professional standards. This discussion generated a number of comments, including the need for a reward structure based on how officers use information provided by crime analysts. Another commented on the fundamental lack of technological understanding within police departments.
The group developed a comprehensive list of 41 potential ways to use predictive policing and how predictive information might be valuable. Read the complete list.
Group members chose to look more closely at vehicle theft/automated early warning system. The group examined how predictive techniques would be used in vehicle theft. Morgan explained that this kind of exercise is useful because departments have the ability to measure their effectiveness. In other words, did predictive policing make a difference?
The Use and Value of Predictive Policing
Participants from the Technical breakout session of the Predictive Policing Symposium discussed the following applications and values of predictive policing:
- Time and location of future incidence in a crime pattern or series
- Identify individuals who are likely to reoffend
- Inmate radicalization risk assessment (i.e., identify inmates who are in danger of being radicalized)
- Drug market displacement (i.e., where next open air drug market will pop up)
- Disorder and environmental variables
- Likely impact of specific operations.
- Disruption of criminal organization (criminal leadership)
- Prediction of criminal adaptation (not only law enforcement efforts but also media, etc.)
- Data analysis and support of crime suppression analysis
- Patrol staffing and resources allocation
- Localized crime spikes
- Identify juveniles likely to be involved in violent crime
- Risk assessment of sex offending in juveniles
- Early identifications of career criminals
- Identify victims of unreported crimes
- Evaluation of interventions
- Impact of drug enforcement on markets and allied crimes
- Identification and analysis of crime-prone events and locations
- Individual-specific analysis
- Travel of serial offenders
- Analysis of predatory patterns
- Correlation of environmental factors outside of crime like weather (i.e., nonobvious correlations)
- Threat and vulnerability assessment
- Prioritization of sources
- Unstructured data extraction (e.g., police reports, blogs, Internet, incident reports and social network analysis)
- Predicting acts of terror
- Predicting riots
- Social network analysis
- Video analytics (including behavioralistics)
- Use of NIBRS to help prediction
- Wide-area surveillance for video fusion
- Precursors and leading indicators to crime (including nonobvious predictors)
- City/neighborhood planning
- Design of spaces; economic development; police, security resource allocation; infrastructure protection
- Offender monitoring, predicting behavior, endpoint sentencing
- Traffic management, crowd control
- Management of police personnel
- Professional development, recruitment
- Risk for excessive use-of-force, discipline
Second Predictive Policing Symposium
At the Second Predictive Policing Symposium, Debra Piehl, president of the Massachusetts Association of Crime Analysts, explained that most small- and medium-size departments struggle with the analytic process, particularly where to start and how much will it cost. Departments must look at the operational impact and efficiency gained from crime analysis. She warned, however, that the only way the process can move forward is by incorporating data quality measures into business practices.
Jim Mallard, technology director of the International Association of Crime Analysts, agreed that data drive everything else. "Analysts must have good, clean data that are readily available," he said. Mallard urged participants to sit down and have a frank discussion about the quality of the data in their data management systems. "The size of the agency does not matter," Mallard continued. "If you have 50 or 500 officers, you can still manage data using programs like Microsoft Access."
Participants identified potential non-traditional data sources like social networking sites. They agreed that the data are limitless, and they need to be creative in what they look for. Susan Smith, a crime analyst in the Shawnee (Kan.) Police Department, encouraged departments to look at new kinds of data such as eviction data, school data, zoning information and mental health data.
Jim Lucht, director of the Providence Plan, discussed one particular tool that agencies can use to access data on returning prisoners. His group worked with NIJ, the Department of Corrections, the Providence Police Department and a reentry service organization to build a system that maps people coming out of prison, offense type and available services based on offense type. "The Internet-based system allows for easy communication between police and probation," Lucht said. The system also enables reentry groups to give service referrals and get statistics from policy initiatives.
Participants also emphasized the importance of having an effective analyst: "If your department is too small, borrow an analyst from a larger neighboring agency," Smith suggested.
"We are changing the way we do police work," explained Michael Freeman, civilian analyst/reserve officer at the Dallas Fusion Center. "It is no longer reactive — it is proactive," he said. "We can throw a dot on a map, but we must know what is behind that dot, what is behind that crime problem." Freeman stressed the importance of having an analyst with a police background. "Crime — analysts who lack a police background often look at the information and do not see that something is missing," he said.