Topics Webinar | EO&GEO Series: GIS Day - GeoAI and the Future of Spatial Analytics

On Wednesday, 20 November 2024, MDPI’s ISPRS International Journal of Geo-Information (IJGI) and Remote Sensing journals organized a webinar entitled "GIS Day - GeoAI and the Future of Spatial Analytics".

This year marks the 25th anniversary of GIS Day, a remarkable milestone in the celebration of geospatial science and technology. To commemorate this occasion, a GIS Day webinar was organized by MDPI’s IJGI and Remote Sensing journals on November 20, 2024, as part of their EO & Geo Series. The theme of the webinar was GeoAI and the future of spatial analytics, highlighting the growing significance of artificial intelligence in geospatial research and applications.

The event drew over 100 registered participants, with approximately 60 attending the webinar live online from across the globe. At the start, Mr. Petar Jeremic welcomed the participants and provided a brief introduction to MDPI, detailing its history, current status, and two of its most prominent open access journals: The ISPRS International Journal of Geo-Information (IJGI) and Remote Sensing. He also outlined the program structure before handing over to Dr. Dev Raj Paudyal, a faculty member from the University of Southern Queensland, who served as the chair of the webinar. Dr. Paudyal provided a brief overview of the webinar’s purpose, elaborating on the significance of the GeoAI theme and the rationale for its selection this year. Furthermore, he introduced three keynote speakers and their respective topics. After the keynote presentations, a Q&A session was organized. Finally, Dr. Paudyal closed the webinar by summarizing the key points from the presentations. Throughout the session, the chat and Q&A box were monitored by Ms. Jing Wen Tan and Mr. Petar Jeremic, ensuring a smooth and interactive experience for all attendees.

The first speaker, Professor Wenwen Li from the School of Geographical Sciences and Urban Planning at Arizona State University, presented on the topic of “Reproducibility and Replicability of Geospatial Artificial Intelligence”, during which she discussed the the science of GeoAI. She introduced three forms of reproducibility (repeatability, reproducibility, and replicability) and explained their relevance to GeoAI. Among these, she highlighted that reproducibility is the most critical aspect for GeoAI research from both computational and spatial perspectives. She also stressed the importance of quantifying the degree of replicability as a key consideration for advancing GeoAI and broader spatial research.

The second speaker, Professor Sisi Zlatanova, head of the GRID Lab and current president of ISPRS TC IV Spatial Information Science, from the University of New South Wales (UNSW), Sydney, Australia, elucidated the power of voxel modeling for 3D spatial data integration and analysis, discussing her own research works. She concluded her presentation by discussing the future research scope in this area such as voxel management (octrees), semantics/attributes, variable resolution, voxel visualization, and parallel processing.

The third speaker, Dr. Yongze Song from Curtin University, Australia, focused on heterogeneity (spatial characteristics) in GeoAI, introducing his latest publications in this area. He started his presentation by addressing a central research question: Which method is the best for spatial prediction? Dr. Song compared and contrasted three models—statistical models, machine learning models, and spatial models—providing insights into their applicability and performance. He concluded his presentation by sharing recent studies conducted by his team, which contribute significantly to the development of geospatial theory and methodologies.

Following the presentations, a lively and insightful discussion ensued, during which the speakers provided additional perspectives, clarified queries, and addressed the critical challenges in GeoAI voxel modeling. Overall, the webinar was highly successful.

The webinar was hosted via Zoom and required prior registration to attend. The full recording can be accessed below. In order to learn about future webinars, you can sign up for our newsletter by clicking “Subscribe” at the top of the page.