Rochisha Shukla, Daniel S. Lawrence, and Bryce E. Peterson February 2020 In 2016, the Urban Institute started working with the Milwaukee Police Department (MPD) to optimize its surveillance system, which consisted of 42 cameras across 40 locations. Improvements included software and hardware upgrades, new high-definition cameras, and two video analytic technologies: automatic license plate recognition (ALPR) and gunshot detection technology (GDT). This brief details the MPD’s experience with these technologies—including benefits and challenges—and provides recommendations for making them more efficient. For public surveillance systems to be effective and efficient, it is important that agencies strategically place cameras where crimes are likely to occur or where evidence is likely to be gleaned (La Vigne et al. 2011; Shukla et al. 2020). The MPD added 45 cameras to its public surveillance network, including 24 panoramic cameras and 12 pan-tilt-zoom (PTZ) cameras. Panoramic cameras are stationary and have wide viewsheds that capture large areas, whereas PTZ cameras allow operators to adjust camera angles and zoom in and out. It is also important for staff to actively and continuously monitor camera feeds (La Vigne et al. 2011), but this can be difficult, even for highly trained camera operators. Video and audio analytics— camera-integrated software that automatically identify persons, objects, and sounds—can aid staff who monitor video recordings or conduct criminal investigations. For these reasons, the MPD added two analytic technologies: PTZ cameras that automatically turn toward the nearest intersection after detecting a nearby GDT alert, and nine ALPR cameras that automatically check license plate numbers against the department’s list of wanted vehicles. During this process, Urban observed operations and interviewed stakeholders to identify benefits and challenges as well as key recommendations for integrating video analytics into police public surveillance networks. JUSTICE POLICY CENTER Lessons Learned Implementing Video Analytics in a Public Surveillance Network