#  2026 CGA Conference Workshops 

 



In conjunction with our 2026 conference, the CGA is pleased to offer five workshops open for attendance by any conference participant who has registered and paid. See workshop information below, and sign up for a workshop using our [**workshop registration form**](https://docs.google.com/forms/d/e/1FAIpQLSdX4lGk_HiU_Y4Upo7K1BjcIinPvindWJq-_EPMJAhPR0Z9Yg/viewform?usp=preview)**.** All workshops will be held the morning of Thursday, October 2 in the CGIS buildings at 1737 and 1730 Cambridge St. Specific rooms for each workshop will be posted soon.

**Workshop 1A: Seeing Green from Space: Machine Learning for Greenspace Detection**

**Instructor:** Martina Pardy, University of Liverpool

**Duration:** 1.5 hours (9:00 a.m. - 10:30)

**Workshop Description:** This hands-on workshop introduces participants to satellite data and machine learning for land use classification using R. Participants will work directly with multispectral satellite imagery, learn to compute vegetation indices such as NDVI, and apply multiple classification techniques, gaining practical, transferable skills in applied spatial machine learning. The workshop speaks directly to the CGA conference's focus on the future of geographic analysis, demonstrating how freely available satellite data combined with open-source machine learning tools can address pressing questions about urban greenspace, environmental inequality, and climate adaptation.

**Workshop 1B: New Approaches in GIS Time Series: Mapping Spatial Patch Transitions via DynamicPATCH and Class Trajectories via QUEST**

**Instructors:** Antonio Fonseca and Aiyin Zhang, Clark University

**Duration:** 1.5 hours (10:30 a.m. - Noon)

**Workshop Description:** Participants will learn to apply two open-source Python packages to summarize land cover trajectories and gross patch dynamics across time series. Unlike traditional metrics that report only net changes, these tools quantify specific temporal alternations and patch transitions. This technical training addresses the CGA conference theme by presenting analytical methods required for impactful geographic research.

**Workshop 2: The WebAI Paradigm of Innovation Research: Extracting Insight From Organizational Web Data Through AI**

**Instructor:** Jan Kinne, ISTARI

**Duration:** 1.5 hours (9:00 a.m. - 10:30)

**Workshop Description:** This workshop introduces the WebAI paradigm, which leverages AI to systematically extract insights on innovation, organizational behavior, and inter-organizational networks from publicly available web data. Participants will learn about the five key properties of organizational web data — vastness, comprehensiveness, timeliness, liveliness, and relationality — and how AI-based methods such as text analysis and hyperlink network modeling can turn this data into actionable intelligence for innovation research and policymaking.

**Workshop 3: Broadening Adoption of Cyberinfrastructure and Geospatial Artificial Intelligence for Disaster Management and Sustainability**

**Instructor:** Zhe Zhang, University of Texas A&M

**Duration:** 3 hours (9:00 a.m. - Noon)

**Workshop Description:** The workshop aims to establish a research network focused on Cyberinfrastructure and GeoAI for blue economy and disaster management research. Participants will gain advanced computational and CI skills, enabling them to apply high-performance GeoAI methodologies to a wide range of disciplines—including Geoscience, Oceanography, Public Health, Engineering, Transportation, and the Social, Behavioral, and Economic Sciences. Leveraging NSF-funded ACCESS and National Artificial Intelligence Research Resources computing resources, the workshop enhances learning and GeoAI skill development through keynote presentations, panel discussions, and training activities.  
This workshop will provide:  
\- Fundamentals of Cyberinfrastructure and High-Performance Computing, including instructions on registering for and accessing NSF ACCESS computing resources  
\- Core concepts in GeoAI for disaster and coastal sustainability management  
\- Techniques for processing and visualizing disaster and coastal science data  
\- Opportunities to engage in collaborative discussions and contribute to papers showcasing the current state-of-the-art in high-performance GeoAI research

**Workshop 4: The UCSB Guide to Teaching Geographic Problem Solving in the Age of AI**

**Instructors:** Peter Kedron, Sarigai Sarigai, University of California, Santa Barbara

**Duration:** 3 hours (9:00 a.m. - Noon)

**Workshop Description:** The rapid development of foundation models and AI agents suggests that automated geographic analysis may soon be a reality. Recent publications have presented foundation models fine-tuned on GIS literature that appear to answer geographic questions with accuracies approaching those of professional analysts. Supported by well-crafted prompts, emerging agentic systems can now decompose geographic problems and construct and execute code. Optimistically, these developments might allow individuals working to solve pressing human and environmental problems to direct their creativity and expertise to the synthesis, evaluation, and reasoning that automation cannot replace. Pessimistically, the same capabilities may lead practitioners to uncritically delegate the guidance and implementation of geographic analysis to AI agents, producing sub-optimal solutions and diminishing the impact of their work. The difference between these outcomes depends substantially on education, that is, on whether students and working practitioners are trained to use these systems to support their most demanding reasoning tasks rather than to displace them.  
  
This three-hour workshop presents a foundation for teaching geographic problem solving with GeoAI, LLMs, and AI agents. The workshop is organized around three interrelated components. First, we identify a set of key AI and geographic concepts and competencies to incorporate into future curricula, courses, and training. Second, we present a pedagogy for teaching the use of these systems, which draws on cognitive load theory, constructivism, and connectivism to help students construct schemas that integrate geographic and AI concepts. Third, we present research results concerning what LLMs and agents are, and are not, well suited to do within geographic analysis, grounding instructional decisions in emerging empirical evidence. Participants will leave this workshop with a set of concepts, candidate schemas, and pedagogical strategies that can be adapted to their own courses and professional training contexts.

**Workshop 5: HeatReady: A Workshop on Geospatial Data and AI for Community Health and Heat Action**

**Instructor:** Andrew Schroeder, CrisisReady

**Duration:** 3 hours (9:00 a.m. - Noon)

**Workshop Description:** The actions required for effective response to the health impacts from extreme heat occur at the local or neighborhood level, but the data that we use to understand the scales and dimensions of those impacts are often generalized based on global sources at the municipal level or in some cases much larger scales. This misalignment between the scale of public health action and the scale of extreme heat analysis led the CrisisReady team to develop a new data API and methodology designed to accelerate, streamline, and localize the science and practice of extreme heat measurement so that cities and health organizations could think more clearly about population needs, program strategies, and the future of urban heat adaptation. By combining downscaled heat measurements with low-cost access to a range of critical high-resolution social and infrastructural data, we can understand a wealth of new opportunities throughout the world for more precise, equitable, and effectively localized heat action.  
  
This workshop invites participants to take up the challenge of local heat and health analysis by working collaboratively to build prototype applications and analyses of neighborhood level dynamics across a diverse range of global cities. Participants will receive a technical overview of HeatReady, the CrisisReady API for downscaled heat, population, and infrastructure data. They will then have the opportunity to work in teams to develop and rapidly prototype new spatial analyses specific to key problems facing local communities based on the resources available through HeatReady.  
  
Examples of key issues to explore include site distribution and resource allocation for adequate cooling, identification of priority areas for building retrofits to reduce heat exposure, cultivation of green corridors for reducing radiant heat in high risk areas, and understanding variances in population vulnerability and health exposure to extreme heat at small spatial scales, and the use of computer vision and street view imagery with downscaled heat data to produce hyperlocalized and predictive understanding of the lived experience of exposure to extreme heat.  
  
While the focus of this workshop will be on the potential applications of the HeatReady API, participants are encouraged to use any additional datasets, tools, AI models, and outside resources needed to enhance their thinking and their work products.

Sign up for a workshop using our [**workshop registration form**](https://docs.google.com/forms/d/e/1FAIpQLSdX4lGk_HiU_Y4Upo7K1BjcIinPvindWJq-_EPMJAhPR0Z9Yg/viewform?usp=preview)**.**