Predictive policing encompasses two broad elements:
1. The analytical techniques that incorporate crime and community data to identify locations at increased risk of crime, likely suspects committing the crimes, and potential victims at risk of being targeted; and
2. The operational implementation of tactics to intervene in or prevent future crimes, solve past crimes and arrest the offenders.
Both elements are vital to a successful predictive policing strategy, with analysis supporting informed operations, and operations providing new data for analysis. The analytical techniques that support predictive policing can be greatly enhanced by the careful selection and implementation of the right analytical software.
More & More Data
Law enforcement agencies routinely collect and store crime and call data in a format that is generally searchable and retrievable for public safety needs such as investigation, prosecution, counseling, legal documentation, officer safety and workload activities. Beyond storage and retrieval, however, most police records management (RMS) and computer-aided dispatch (CAD) systems are not fashioned to service analytical needs. Consequently, an additional software or data system is required for statistical or predictive analysis.
In smaller, low-call-volume agencies, standard office software may be used to categorize and organize data to conduct effective analysis. However, the amount of data that is collected and available is growing for a number of reasons, regardless of the size of the population being served. Police data collection is moving away from paper documentation and hand-written logs, allowing for more detail in police reports and automated activity tracking. Improvements in computer processing speeds and data storage volume allow for larger and more complicated datasets with more data fields and record counts. And besides just collecting and storing more internal agency data, high-speed data connections allow for disparate data sources to increasingly become available for fusion and analysis.
The bottom line: Agencies of any size would benefit from a robust analytical software package capable of organizing and fusing these information-rich datasets to support predictive policing.
Selection Criteria
When confronted with the myriad software packages, individual mapping and statistical tools, proprietary RMS and CAD systems, and impressive marketing claims, the selection process for a sound and effective analytical software solution for predictive policing can be cumbersome and daunting. There are important considerations for data integrity and security, data translation between existing systems, real- or near-time updates, training and support and an understanding of relevant analytical techniques to support the predictive policing mission.
To help sort through the options, focus on the following criteria.
Security
The data system must be secured to protect sensitive law enforcement information, including open investigations, victim identities and personally identifiable information (PII). There must be controlled access through individual logins and secure password and data encryption. Systems that are CJIS-compliant have already addressed these data security issues according to the FBI's Criminal Justice Information Services (CJIS) Security Policy.
System compatibility
A data-management solution should provide translation between the agency's existing CAD and RMS systems as well as other disparate data sources to establish a common schema to interact with the data in a meaningful way.
Live data & automation
The data system should be automated to update regularly, in real-time or near-real-time, from the CAD and RMS data that is under the control of the police agency to ensure that users always have the most recent relevant information possible.
Ease of use
The interface should be intuitive enough for casual users to navigate the interactive maps and set up user-defined charts or "dashboards" based on real-time data, but customizable and robust enough for the crime analyst or other "power users" to make full use of more advanced tools. Editor's note: Agencies should assess the capabilities of their staff to ensure that acquired tools can be effectively used. Agencies without a dedicated analyst will want to look for programs that are easy for new users and that offer training, tutorials or online assistance.
Relevant toolsets
An analytical software solution should have a variety of capabilities relevant to the agency's predictive policing mission. These may include interactive crime maps available online or on an intranet, live-data access from multiple sources, automated and ad-hoc reports, regional data-sharing capabilities, mathematically sound temporal and spatial analysis tools and predictive analytics based on industry-recognized research and best practices. Analytical software that has only one primary function or algorithm, such as density mapping, near-repeat calculation, linear regression forecasting, geographic profiling, temporal topology, or risk-terrain modeling, should be considered useful for the tool that it is, rather than viewed as a complete analytical solution.
Data sharing & collaboration
Improved interagency cooperation, as well as the rise of fusion centers and task forces, are enabling agencies to increasingly collaborate on cross-jurisdictional crime, organized crime and federal problem sets at a local level. The data system should have the capability to support regional data-sharing and analysis through secure individual logins and agency-controlled access permissions.
Price
The up-front and annual maintenance fees for an analytical software solution should be affordable and predictable, ensuring that the agency continues to receive high value to meet their analytical needs. Budgeting for updates that should include analytical and technological capabilities will allow the agency to maximize the effectiveness of this software investment going into the future.
Training & support
In addition to effective initial training for all users and system administrators in the agency, a robust analytical software solution should offer ongoing technical and analytical support to ensure that the system remains up-to-date and operationally relevant to the agency's predictive policing mission.
Tap the Experts
When selecting predictive policing software, remember that the crime analyst is a great asset in vetting the right solution for the agency. Through familiarity with the datasets in the context of agency operations, knowledge of appropriate analytical techniques, training in various analysis-related software programs, better-than-average technical skill with computers, and, as a primary operator, a vested interest in acquiring the most effective software to meet the agency's needs, the analyst can help guide the agency to a reasoned decision.
