Modelling state-level crash fatalities by a network-constrained inhomogeneous poisson point process

Traditional spatial analysis of traffic crashes often relies on planar density estimation, which ignores the topological constraints of the road network and limits the precision of risk assessment. To address this limitation, this study employs a Network-Constrained Inhomogeneous Poisson Point Process (IPPP) to model fatal crash risk across the diverse roadway environments of South Carolina. The model integrates high-resolution road geometry (slope, curvature) with dynamic traffic and environmental conditions. To account for exposure heterogeneity, we explicitly normalize fatal incident intensity by total crash frequency, enabling the estimation of the conditional risk of fatality given a crash occurrence. A semi-parametric Generalized Additive Model (GAM) is further incorporated to validate the robustness of the model's parametric specification. Additionally, we conduct comparative analyses between highways and local roads to capture context-specific risk mechanisms. Results reveal a distinct divergence in risk mechanisms: while higher traffic volumes (AADT) exhibit a strong protective effect on conditional fatality rates, steep gradients and commercial vehicle involvement act as universal risk amplifiers. Furthermore, the analysis uncovers spatial heterogeneity in geometric hazards: unlike slope, the lethality of sharp curvature is highly context-dependent, posing a significant threat primarily on local road networks. These findings demonstrate the capability of the proposed network-constrained framework in identifying micro-scale geometric determinants at the state level, providing actionable insights for precision safety planning

Crash · Point (geometry) · Point process · Poisson distribution · Poisson process · Process (computing) · Gene Regulatory Network Analysis · Reliability and Maintenance Optimization · Simulation Techniques and Applications

• A multilevel regression model to explain disparities in residential cooling infrastructure across heat-threatened regions

  Open Access•Yuting Yang, Yang Yang et al.•Applied Geography•2026

  Add to list

• The Modifiable Areal Unit Problem in Multivariate Statistical Analysis

  Open Access•A Stewart Fotheringham, D W S Wong et al.•Environment and Planning A…•1991

  Add to list