This website is a companion to the paper “Log-Gaussian Cox
Processes on General Metric Graphs” by David Bolin, Damilya
Saduakhas, and Alexandre B. Simas. The paper is available on arXiv, and all code is
provided in the GitHub
repository.
Overview
This website provides code and extended examples to reproduce
the traffic accident analysis from the paper. While raw data
cannot be shared, the methodology is demonstrated comprehensively.
Key Goals
- Reproduce Results: Step-by-step code for fitting
LGCPs to traffic accident data.
- Methodology Guide: Adapt the framework to similar
problems (covariates, graph construction).
- Reproducibility: Templates for inference, and
visualization.
Navigation Guide
Use the navbar above to explore the workflow:
- Graph Setup: Build metric graphs from
road networks.
- Covariates: Prepare traffic/road
covariates (speed limits, road type).
- LGCP Models: Fit Model 1 (spatial effects)
and Model 2 (spatial + covariates).
- Risk Analysis: Compute exceedance
probabilities for high-risk roads.
- Interactive Maps: Visualize results
spatially.
Citation
Bolin, D., Saduakhas, D., & Simas, A. B. (2025). Log-Gaussian
Cox Processes on General Metric Graphs. arXiv:2501.18558.
For questions or feedback, visit the GitHub repository or
contact the authors.