Revealing Hurricane Harvey 2017

Geovisualization of twitter-based trajectories; on Esri Map Book Vol.39

View Full Image

Social media platforms such as Twitter with user-generated contents provide great potentials to better understand the spatiotemporal and transportation pattern of evacuation behaviors during hurricanes. Hurricane Harvey was the costliest tropical cyclone on record (tied with Hurricane Katrina of 2005), inflicting roughly $125 billion in damage across the Houston metropolitan area and Southeast Texas. It lasted from about mid-Aug until early Sep 2017, with many records for rainfall and landfall intensity set during that time.

The overall map visualizes the trajectories generated based on a Twitter dataset relevant to 2017 Hurricane Harvey. The transportation mode (flight vs. drive) is calculated using machine learning algorithms considering multiple spatiotemporal and semantic factors, based on manually classified ground truth dataset. Technical details of the transportation mode detection process are introduced in the corresponding research paper. The four stage-specific maps display the aggregated trajectories at CBSA (core-based statistical area) level during the 4 stages of Hurricane Harvey correspondingly, indicating the connections among major metropolitan areas. Both the line width and opacity represent the volume of connections between two cities. Only strong connections are displayed, while the weaker ones are hidden.

This map is selected to be published on Esri Map Book Volume 39 (2024). The initial version is presented at North American Cartographic Information Society (NACIS) 2021 Student Map Gallery. The updated version is displayed at Map Gallery of Big Ten Academic Alliance (BTAA) Geospatial Information Network (GIN) 2023 Conference.

Below are examples of original maps in research paper which this comprehensive map is based on.