A GeoAI Framework for Predictive Risk Monitoring of Electrical Substations and Energy Infrastructure
Olalekan Alamoh, Georgia State University
This presentation introduces a geospatial artificial intelligence (GeoAI) framework designed to support predictive monitoring of electrical substations and critical energy infrastructure. The approach integrates satellite imagery, environmental data, and spatial analysis to identify risks such as flooding, extreme heat, and vegetation encroachment before failures occur. By combining multiple data sources into a unified system, the framework enables proactive decision-making and enhances infrastructure resilience. The presentation will highlight system design, data integration, and visualization through an interactive dashboard, demonstrating how GeoAI can transform infrastructure management from reactive response to early risk detection.
Maps going viral: learnings across public health cartography
Eric Brelsford, Stamen Design
Kelsey Taylor, Stamen Design
Stamen has collaborated with several universities and NGOs in recent years on health-related cartography projects, including WastewaterSCAN, the UCSF Health Atlas, the Duke SEED Atlas, and more. By engaging directly with epidemiologists on these projects, we have had the freedom to explore new concepts and ideas with the confidence that experts will ensure we are telling the right stories and portraying ongoing outbreaks accurately.
Through these examples, this talk will focus on patterns in the health mapping space, how it differs from other types of cartography that we work on at Stamen, and what we can apply to future public health-related projects.
MadTrails: A public land management and trail accessibility tool
Susannah C. Cox, University of Wisconsin-Madison
Alton L. Hipps, University of Wisconsin–Madison
Natalie Correa, University of Wisconsin–Madison
Jonathan K. Nelson, University of Wisconsin–Madison
We discuss the tool MadTrails, a multi-application suite built on the ESRI ArcGIS Online environment for the University of Wisconsin–Madison Arboretum. Built alongside community stakeholders, MadTrails allows for online reporting of maintenance and access issues by Arboretum volunteers and the public to aid in better management of the arboretum space and increased access to communities with mobility difficulties. Additionally, we include opportunities for direct mapping of accessibility issues. In this talk, we will discuss the functionalities of MadTrails, the stakeholder engagement process, initial evaluation findings, and future steps for the application.
Communicating Healthcare Affordability Through Maps: A Visual Analytics Framework Using NHIS and MEPS Data
Sandipta Khare, Claremont Graduate University, Ph.D. Student, Information Systems and Technology
Alexis Parker, Claremont Graduate University
Prof. Hengwei Zhang, Claremont Graduate University
Healthcare affordability can be understood through both barriers to obtaining care and financial burden after receiving care. This study uses two national public-use survey datasets, the 2023 National Health Interview Survey (NHIS) and the 2023 Medical Expenditure Panel Survey (MEPS), to examine these related dimensions. NHIS is used to analyze reported cost-related barriers to mental health counseling and therapy, while MEPS is used to examine annual healthcare expenditure burden. Through regional mapping, visual analytics, and uncertainty-aware cartographic design, the project demonstrates how public-use health survey data can be used to communicate healthcare affordability patterns responsibly while acknowledging geographic limitations.
Network Ontologies for Disaster Management Tools: A Review of Cartographic Literature
David Retchless, Texas A&M University at Galveston
Carolyn Fish, University of Oregon
Jim Thatcher, Oregon State University
Chad Giusti, Oregon State University
Conor Progin, Oregon State University
Courtney Thatcher
We review cartographic literature for the network ontologies of disaster management tools, identifying gaps and opportunities. Natural disasters disrupt flows of people, energy, information, and other resources that are essential to modern systems. Cartographic tools for disaster management therefore frequently model and map these flows and associated networks. Understanding the network ontologies (implicit or explicit) that undergird these tools is essential to identifying their capacities and limitations. Differences in what counts as a node or edge – what features, with what dimensionality and level of attribution – can be the difference between delivering life-saving insights into network function and missing key vulnerabilities.
