Modeling of Pedestrian Activity at Signalized Intersections: Land Use, Urban Form, Weather, and Spatiotemporal Patterns

Abstract

The present study evaluated the effects of land use, urban form patterns, and weather conditions on pedestrian activities. In this research, 8 h of daily manual pedestrian counts during the a.m. peak, noon, and p.m. peak periods were collected from a large sample of signalized intersections throughout Canada. The use of different model settings was attempted as part of a model sensitivity analysis. Results revealed that a 100% increase in population density or the amount of commercial space around intersections increased pedestrian flows by 22.7% to 37.1% and 10.7% to 11.7%, respectively. Moreover, the pedestrian activity at an intersection decreased as the distance from downtown increased (with an elasticity of 44%). Also, very warm weather (with temperatures >30°C) decreased pedestrian activity by up to 22%. This study was the first attempt to develop a spatiotemporal model of pedestrian activity in a large city. These results should be taken with caution, however, because the sample of intersections was not randomly selected. Moreover, the modeling techniques used in this research did not take into account the potential spatial correlation across intersections. This factor may have caused bias in parameter estimates. These issues are part of future work.