Predictive Policing with the Help of Analytics

Disclaimer This piece was written by the author as a white paper assignment for a graduate studies program.

The brave men and women that take the oath to serve as police officers have always faced a number of challenges on the job: violent criminals, drug offenders, domestic violence situations, young children making dumb mistakes, even adults making the same mistakes, and many more. Each of these situations is laden with multifaceted issues that contribute to these behaviors and must be considered when making decisions that will directly impact the lives of the officer, the offenders themselves and ultimately the community at large. As society barrels towards the impeding technological revolution, police officers have evolved along side the rest of us and implemented the use of modern tools such as digital software and systems, portable body cameras, and more recently, predictive analytics. This paper seeks to explore the world of predictive analytics in policing by assessing the impact it has had on the law enforcement community of Memphis, Tennessee, since the implementation of its Blue C.R.U.S.H. program in 2006, discuss the findings and their result on departmental decisions, and considering the possible ethical components associated with its implementation.

Blue C.R.U.S.H. (Crime Reduction Utilizing Statistical History) brought a refreshing spin on the business of policing in the city of Memphis, Tennessee. The beginning of the 21st century was frightening for Memphis, having seen crime on the rise for nearly a decade and eventually landing the reputation as “one of the most crime-ridden cities in the United States,” in 2005 (Piyanka Jain & Puneet Sharam, 2015) . In an attempt to eradicate this issue, local college professor Richard Janikowski teamed up with the Memphis Police Department to create Blue C.R.U.S.H., a statistical algorithm that converted raw data collected from past crimes into usable information. While the police department had been collecting information on the location and the nature of crimes for years, Janikowski helped them to use the information to their advantage by helping the department create maps that could pinpoint the possible future crimes based on factors such as time, date and location. By the year 2012, crime rates were down across the board and these predictive practices had become commonplace, being used on a daily basis by many of the officers in their work. In the eyes of the department heads, and many of the citizens, the program was wildly successful and was even emulated at police departments across the country.

Six years after implementation, the economic benefits became obvious. In the year 2013 violent crimes were down by 26 percent and car-theft, muggings, and petty crimes were decreased by nearly 40 percent in each category (James Vlahos, 2012). Such numbers are enough to impress even the strongest of skeptics, but more importantly to decision makers, the police department had realized a return on investment (ROI) of 863 percent after just 6 years of implementation based on the number of officers necessary to achieve such impressive rates of crime reduction in the same period of time (Piyanka Jain & Puneet Sharam, 2015) .

From a business standpoint, this police department was making all of the right moves: crime was down, the city was saving money both on officer salaries and resource deployment, and the city was becoming a more attractive place to live. In fact, one study showed that violent crimes in apartments were down more than 35 percent, which is extremely impactful when you consider the rate of renter-occupied housing in Memphis was more than half of the population at the time (Walter L. Perry, Brian McInnis, Carter C. Price, Susan C. Smith, & John S. Hollywood, 2013) . In Memphis, the majority of apartments are occupied by individuals categorized as low income with volatile family situations, both indicators for potential crime risks (“Quick Facts,” n.d.) . Many of the “hotspots” correlated with low-income apartment housing, and were commonly visited by police on regular patrol (James Vlahos, 2012). The positive results of this predictive policing strategy were only made possible by the correct application of the data itself. Though the Memphis PD had been collecting the information about crimes around the city for years, it was not a useful source of information until they figured out a way to turn the raw data into insights, resulting in a positive impact on their financial bottom line and ultimately improving the efficiency and functionality of the entire department. Experts are quick to assert that analytics are only as powerful as the impact they drive, and in the case of Memphis the impact was substantial.

While the statistics discussed earlier are impressive, it was the actions taken by leaders in the department as a result of the analytics that made the real difference. Lead decision makers were able to utilize the focus area maps that were created in a way that allowed them to deploy more effective resources by making sure the appropriate task force was available to respond to certain situations, saving time and money across all departments (Piyanka Jain & Puneet Sharam, 2015) . Having the ability to deploy a trained task force such as the drug prevention unit to a time and place of likely drug activity proved highly effective compared to having a general beat cop in the area because these officers were more in tune with the actions and behaviors of drug offenders, making their time on the streets more effective. The innovative policing strategies utilized by the Memphis Police Department were driven by accountability further strengthening the reputation of the Blue C.R.U.S.H. program. Moreover, leaders at Memphis P.D. took proactive steps to ensure accuracy by holding meetings centered around tracking for responsibility, accountability and credibility of the system (James Vlahos, 2012) . It was the continual tracking, measurement, and adaptions based on results from Blue C.R.U.S.H. that insured the continued success of the program. These and other actions taken improved community relations and aided in the creation of effective partnerships including many innovative programs that are still active today. One such program is the Community Outreach

Program (C.O.P.), which focuses on the reduction of juvenile violence. It is clear that the actions taken by leaders at the Memphis Police Department based on the analytical results of Blue C.R.U.S.H. created a direct impact on crime rates in the city. While the data suggests all positive outcomes from the program, it is necessary to review the ethical components associated with predictive policing. Profiling, or the act or process of extrapolating information from a person based on known traits or tendencies as defined by Merriam-Webster, has been a hot button topic in recent history. As has been proven on more than one occasion, police profiling can quickly lead to incorrect assumptions that can cause unconstitutional violations of rights, and in the worst of cases, unnecessary death. When considering the intricacies of predictive policing the question of whether of not individual freedoms are being infringed upon must be asked. While the current system is careful not to individualize data, it does have the propensity to categorize individuals with particular backgrounds as “more likely” to commit a crime regardless of actual behavior (Zwitter, 2014).

The real danger comes into play when an individual is targeted and prematurely deemed guilty of a crime based on association.While this has yet to be seen in Memphis, there are areas of predictive analytics in which this individualized practice has been seen, like with the work of Richard Berk, a professor of Criminology at the University of Pennsylvania. Berk has created and algorithm that can predict the likely-hood of someone on parole or probation to commit murder within a 75 percent accuracy. Going a step further, some parole boards are using Berk’s work as a tool for consideration in regards to releasing violent criminals (James Vlahos, 2012, p.). When applied in this way, an ethical question of the ages is formed: is a 25 percent margin of error accurate enough to grant or deny individual freedom? While it is a useful tool, and provides more insight than the previous practice of gut instinct, it is important that the human piece of the equation remains fully intact when making parole decisions. Experts in the subject are quick to disassociate the Blue C.R.U.S.H. program with this type of behavior, affirming that the Memphis tactics are focused on the occurrence of actual crimes based on a particular time or place, but some critics are skeptical that there is a very small step from one to the other. It is clear to see that technology is becoming commonplace in police departments across the nation, with big data tactics being a radical disruptor in the way officers approach situations on the daily.

Predictive analytics in policing poses positive and negative characteristics based on their application, though this paper found current practices to lean much more towards the positive. Moving forward, Memphis, and many other police departments for that matter, must use extreme caution when employing these tactics by ensuring that implementation of strategies continues to be driven by consistency, accuracy, and accountability when reaping the highly impactful benefits of these practices. If done correctly, predictive analytics could reshape the world of police work, making it more effective, less expensive, and provide safer communities for citizens across the country.

Bibliography

Piyanka Jain, & Puneet Sharam. (2015). Behind Every Good Decision: How Anyone Can Use

Business Analytics to Turn Data into Profitable Insight. New York: AMACOM.

Quick Facts: Resident Demographics| NMHC.org. (n.d.). Retrieved November 9, 2017, from

http://www.nmhc.org/Content.aspx?id=4708

James Vlahos. (2012). The Department of Pre-Crime. Scientific American, 306(1), 62–67.

Walter L. Perry, Brian McInnis, Carter C. Price, Susan C. Smith, & John S. Hollywood. (2013).

Predictive Policing: The Role of Forecasting in Law Enforcement Operations.

Washington D.C.: RAND Corporation.

Zwitter, A. (2014). Big Data ethics. Big Data & Society, 1(2), 2053951714559253.