Workshop

One-day Workshop on Spatiotemporal Innovation and GeoAI Applications

Time: July 22, 2024 (Monday), one day before the Symposium on Spatiotemporal Data Science (July 23-24, 2024)

Venue: Virginia Tech Research Center – Arlington, 900 N Glebe Rd, Arlington, VA 22203

Organized by

  • The Spatiotemporal Innovation Institute Program

Description of the Workshop

This workshop aims to provide a comprehensive overview of the potential and challenges of some cutting-edge technologies in advancing geospatial science and applications, fostering a deeper understanding and encouraging further exploration and innovation in the field. The workshop will introduce some new tools and applications of GeoAI built on the workflow technology as well as their applications in social science and public health. Participants will learn the GeoAI-based methodology for geospatial analysis, GeoAI tools and packages for spatial data analysis, as well as their applications.

Topics: 6 hours

  1. Replicable Data Analysis with Geospatial Analytics for KNIME
  2. Develop GeoAI Tools using ChatGPT
  3. Cloud Computing with Google Earth Engine and GeoAI
  4. Geospatial Methods and Tools for the Spatial Assessment of Healthcare Accessibility

Workshop Registration here

Audience: Anyone who is interested in learning the essentials of replicable workflow technology, GeoAI methods and applications in cross-disciplinary domains.

Abstracts and Instructors:

I. Replicable Data Analysis with Geospatial Analytics for KNIME

This workshop will introduce the recent development of workflow technology for replicable spatial data analysis. The topics include: (1) Introduction to KNIME, a free tool for workflow data analysis; (2) Introduction to Geospatial Analytics Extension for KNIME; (3) GeoAI data analysis with KNIME; (4) Case studies of GeoAI and KNIME applications for environmental and socioeconomic studies with big data.

Instructor:

  • Lingbo Liu, Center of Geographic Analysis, Harvard University

II. Develop GeoAI Tools using ChatGPT and Python packages

The first half of this session will introduce GeoLocator – a GeoAI tool re-developed from ChatGPT to detect the location based on the image that was input into the ChatGPT. It will more broadly introduce how to reformulate ChatGPT for geographic studies and research. The second half of this session will introduce a few advanced geospatial and GeoAI methods for spatial prediction, including geographical detectors model, geographically optimal similarity model, the second-dimension spatial association model, and popular R packages which have been downloaded over 120,000 times globally. Case studies of using these methods and tools will be introduced in the session.

Instructors:

  • Siqin Wang, Spatial Sciences Institute, University of Southern California
  • Yongze Song, School of Design and the Built Environment, Curtin University

III. Cloud Computing with Google Earth Engine and GeoAI

This workshop explores the integration and applications of cloud computing technologies, specifically Google Earth Engine, with Geographic Artificial Intelligence (GeoAI) to address complex spatial problems. The presentation aims to showcase how cloud computing offers scalable and efficient computing resources for processing vast amounts of geographic data, enabling researchers, scientists, and developers to perform advanced spatial analysis and machine learning tasks without the need for extensive hardware infrastructure. Case studies or examples are provided to illustrate the practical applications of combining Google Earth Engine and GeoAI.

Instructors:

  • Xiao Huang, Department of Environmental Sciences, Emory University
  • Qiusheng Wu, Department of Geography & Sustainability, University of Tennessee

IV. Geospatial Methods and Tools for the Spatial Assessment of Healthcare Accessibility

The workshop will introduce some recent development of methodology and technology for the spatial study of health accessibility. The topics include: (1) Overview of the issues on spatial accessibility, (2) Two-Step Floating Catchment Area (2SFCA) Method, (3) Generalized 2SFCA (G2SFCA), (4) Inverted 2SFCA (i2SFCA) method for estimating potential crowdedness in facilities, (5) Two-Step Virtual Catchment Area (2SVCA) method for measuring accessibility via internet or virtual accessibility, and (6) the ArcGIS toolkit and KNIME workflows for automated implementation of a case study of various accessibility measures for primary care physicians in Baton Rouge metropolitan region.

Instructors:

  • Fahui Wang, Graduate School, Louisiana State University
  • Changzhen Wang, Department of Geography, University of Alabama
  • Mengxi Zhang, Carilion School of Medicine, Virginia Tech