Dynamic Compositional Modeling of Pedestrian Crash Counts on Urban Roads in Connecticut

• Activity Number: 398
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Volodymyr Serhiyenko*+ and John Ivan and Nalini Ravishanker and Md Saidul Islam
• Companies: University of Connecticut and University of Connecticut and University of Connecticut and University of Connecticut
• Keywords: Compositional time series ; Box-Cox transformation ; Dynamic Generalized Linear Models ; Injury severity levels ; Pedestrian safety ; Vector autoregressive (VAR) models

Abstract

Uncovering the temporal trend in crash counts provides a good understanding of the context for pedestrian safety. This paper describes statistical methods uncovering patterns in monthly pedestrian crashes aggregated on urban roads in Connecticut from January 1995 to December 2009. We investigate the temporal behavior of injury severity levels, as proportions of all pedestrian crashes in each month, taking into consideration effects of time trend, seasonal variations and VMT (vehicle miles traveled). We describe a dynamic framework with vector autoregressions (VAR) for modeling and predicting compositional time series. Combining these predictions with predictions from a univariate statistical model for total crash counts will then enable us to predict pedestrian crash counts with the different injury severity levels. We implement the Integrated Nested Laplace Approximation (INLA) approach to enable fast Bayesian posterior computation. Taking the least severe injury level as a baseline we conclude that there was a noticeable shift in the proportion to the less severe injury level suggests that the overall safety on urban roads in Connecticut is improving.