MetricsVis: A visual analytics framework for performance evaluation of law enforcement officers

Abstract

Limited resources and increasing costs require law enforcement agencies to develop effective methods for measuring and evaluating officer performance. The methods enable law enforcement to be more effective in their event planning, resource allocation, decision making, and community policing efforts. The paper introduces a visual analytics framework for efficient measurement and evaluation of officers' performance through interactive and coordinated visual dialogs. Through collaboration with our partner law enforcement agency, we have developed a comprehensive categorization of offense types utilizing a crowdsourcing approach. Our system allows end-users to interactively specify the offense types and customize the performance metric based on their domain knowledge and policing priorities. The performance scores for each officer are then visualized based on a matrix representation. The representation supports a rich set of interactions including selection, filtering, ranking and correlation to allow end-users to supervise and refine the performance evaluation process. With our system, end-users can explore the activity patterns and performance trends for either a large group or an individual, and identify critical factors that help to improve the operational decision making process. To demonstrate the proposed approach, we present two case studies and provide domain expert feedback.