Before/After studies of safety surrogates can focus on a host KPIs. The following metrics will help the City of Atlanta demonstrate project effectiveness without relying on crash data alone. ​ Time-To-Collision (TTC) is the time remaining before the paths of two crossing road users would intersect (and a crash would occur) if their directions and speeds are maintained. Smaller TTC values indicate a collision is more imminent between two road users, while larger values indicate that there is more opportunity for one or both users to alter their paths and/or speeds to avoid a collision. Larger TTC values can indicate that the location is experiencing fewer close calls. ​​ Post-Encroachment Time (PET) is the time from when one road user leaves a specific location (or conflict point) to the time another user arrives at that same location. Smaller PET values indicate a higher likelihood for a collision, and an increase in average PETs can indicate that the location is experiencing fewer close calls. ​​ Red Light Running is assessed by instances of vehicles entering the intersection despite not having a permissive green or yellow light. Fewer instances of red light running may indicate fewer conflicts that may result in a severe crash. ​​ Crossing on Opposing Green is assessed by instances of pedestrians crossing the street without a green light or walk signal. It may indicate that signal times need to be adjusted so pedestrians can cross in an efficient manner and do not feel compelled to risk crossing against the signal. ​​ Typical Speeds are captured and instances of speeds 10 mph over the speed limit are evaluated. Average speeds are also calculated. Slower speeds increase potential reaction time for drivers, as well as reduce the force of impact if a crash occurs. ​​ Counts of Vehicles, Pedestrians, and Cyclists are captured as they enter a conflict zone (e.g., intersection). Counts of users at an intersection, intersection approach, or midblock section of road provide important information on exposure and data for normalizing the results.

Video assessments are conducted at each location. They can last 24-72 hours depending on the circumstances at each location. The extended timeframe allows the City to capture operations at the peak times and days, as well as periods of lower activity. A proprietary algorithm detects potential conflicts and safety issues, capturing data for to calculate the KPIs. ​ The City can also capture road user counts at each location (such as vehicles, pedestrians, and cyclists) to demonstrate project usage. A safer system encourages a more efficient and active road network, and these before/after evaluations can help show that safer infrastructure benefits everyone.