Evacuation Traffic Prediction for Dynamic Transportation Networks over Multiple Hurricanes: A Generalized Graph Learning Approach

Source
DataCite
Published
2026
Country
United States
License
Open Data Commons Attribution
Themes
Disaster Risk & Resilience, Sustainable Mobility, Natural Disasters

Get the data from DataCite

About this dataset

This project contains the dataset used for the study titled “Evacuation Traffic Prediction for Dynamic Transportation Networks over Multiple Hurricanes: A Generalized Graph Learning Approach.” The study develops a Dynamic Graph Learning (DGL) framework to predict traffic conditions during unseen hurricane evacuation periods. The project focuses on evacuation traffic flow prediction using detector-based traffic data, dynamic transportation network representation, evacuation order information, and spatial attributes related to hurricane events. It also introduces a modular training approach to show the incremental learning of the DGL framework from multiple hurricanes. The dataset includes processed traffic detector (MVDS) observations from multiple historical hurricanes in Florida, along with multiple evacuation-related features used for model training and evaluation. The dynamic...

How the data was collected

The dynamic graph structure represents changes in the active transportation sensor network over time, allowing the model to account for real-time detector availability and evolving roadway conditions during evacuation events.

Browse sustainability datasets on SusanHub