Austin Vision Zero Study
TIME-SERIES FORECAST
In Austin, we conducted a comprehensive machine learning (ML) time series forecasting analysis, spanning approximately 175 census tracts. Each tract was analyzed to predict the number of traffic crashes per day over a 14-day period. To visually communicate these predictions, we utilized a color gradient on our map visualization. Darker shades indicate tracts with a higher likelihood of experiencing three or more crashes per day, while lighter shades signify areas with fewer expected crashes.
This visual tool is essential for quickly identifying and prioritizing high-risk areas within the city. By pinpointing these zones, local authorities and city planners can deploy targeted interventions to address specific safety challenges. These interventions may include enhanced traffic monitoring, improved road signage, and community awareness programs tailored to reduce crash rates effectively.