Transforming Law Enforcement with Precision and Data

Data-Driven Insights for Effective Vehicle Recovery

In a conference room at the Portland Police Bureau’s east precinct, former software developer Michael Terrett introduced officers to a new method of identifying stolen vehicles. Using a deck of cards, he demonstrated how probabilistic thinking could enhance their efforts. Officers from across the metro area gathered to learn this novel approach, which treats each vehicle detail as a clue to its legitimacy.

Terrett’s initiative marked a significant shift in police tactics. Instead of relying solely on basic vehicle information, the bureau began tracking granular details like broken windows, tinted glass, and erratic driving behavior. This approach allows officers to calculate the likelihood of a vehicle being stolen based on multiple factors, leading to more targeted and effective interventions.

From Hypothesis to Reality: The Impact of Detailed Data Collection

The results have been striking. In 2022, Portland reported over 10,900 stolen vehicles, breaking a three-decade-old record. However, by mid-2023, that number had plummeted to just over 4,200. The dramatic reduction can be attributed to the bureau’s meticulous data collection and analysis methods.

Terrett’s widget, displayed on patrol car screens, calculates the probability of a vehicle being stolen in real-time. Each factor—such as missing license plates or spray-painted exteriors—adjusts the likelihood score. For instance, a Toyota pickup initially scored 4%, but adding details like "missing plates" increased the score to 17%. By focusing on high-probability vehicles, officers achieve better outcomes with fewer stops.

Field Application: Putting Theory into Practice

Gresham police officer Jack Labuhn exemplified the practical application of this data-driven approach during a mission near the Portland-Gresham line. Spotting a lifted Ford pickup without license plates, Labuhn and his partner initiated a pursuit. The truck’s erratic behavior, including an abrupt turn into a gas station without diesel pumps, raised suspicions. After reinforcements arrived, the suspects were apprehended, and the truck was recovered.

Inside the vehicle, evidence of tampering was evident: cut ignition wires, piled clothes, and burn marks on the door handle. These findings were catalogued into the bureau’s dataset, further refining the criteria for identifying stolen vehicles. Labuhn emphasized that this method has become integral to his work, noting that it could benefit law enforcement agencies nationwide.

Training and Adaptation: A New Era of Policing

Before joining the force, Terrett spent years in the tech industry, where he developed a keen appreciation for data. He brought this expertise to the bureau, convincing leadership to adopt a more analytical approach. Initially, seasoned officers relied on gut feelings and experience to identify stolen vehicles. Now, data provides concrete evidence, enabling newer officers to quickly develop precision in their assessments.

This shift not only improves efficiency but also ensures fairness. By focusing on high-probability targets, officers limit unnecessary stops, reducing the impact on innocent drivers. The strategy has garnered national attention, with presentations to Interpol and the U.S. Department of Justice, highlighting its potential for broader adoption.

Challenges and Future Directions

Despite its success, the program faces ongoing challenges. Factors like evolving criminal techniques and legal changes can influence outcomes. For example, manufacturers have issued recalls for vulnerable models, while court rulings may alter arrest protocols. Nevertheless, the data-driven approach remains adaptable, continuously improving through feedback and analysis.

Researchers like Dr. Jeff Tyner from OHSU’s Knight Cancer Research Institute see parallels between this policing method and cancer research. Both fields involve adapting to changing conditions and optimizing strategies based on data trends. As crime patterns evolve, the bureau’s ability to quickly adjust will be crucial in maintaining its effectiveness.

Beyond Auto Theft: Broader Implications for Public Safety

The impact of this data-driven approach extends beyond auto theft. Recovered vehicles often contain evidence of other crimes, such as drug possession, burglary, or identity theft. This comprehensive approach enhances public safety by disrupting criminal networks and providing valuable intelligence to law enforcement.

Ultimately, the Portland Police Bureau’s innovative strategy demonstrates the power of data in transforming traditional policing methods. By combining technology, training, and adaptability, they have set a new standard for effective crime prevention and recovery.