Conference on Environmental Science and Technology

Air pollution is a critical global health concern, contributing to millions of premature deaths each year. Although significant progress has been made in air quality monitoring, there remains a gap in predictive tools that effectively link pollution exposure to health outcomes. To bridge this gap, we developed ODESSA, an AI-powered WebGIS platform designed to forecast hospital admissions attributable to air pollution. ODESSA integrates diverse data sources, including air quality measurements, satellite imagery, meteorological variables, demographic information, and health records, to generate spatially explicit risk assessments. These insights aim to support proactive public health planning and response. Our approach employs advanced machine learning techniques, to uncover patterns and correlations between air pollution and hospitalizations. These models are trained on historical datasets to capture both short-term pollution spikes and long-term exposure effects. Preliminary results from pilot studies conducted in Lisbon, Portugal, demonstrate promising predictive accuracy, offering a valuable tool for healthcare systems to anticipate and manage patient surges. An interactive WebGIS interface is currently under development to enable policymakers, urban planners, and health authorities to visualize high-risk areas, evaluate intervention strategies, and optimize the allocation of healthcare resources. This work aligns with the Sustainability and AI track by showcasing how artificial intelligence can enhance environmental health management. By combining AI and spatial analysis, ODESSA offers real-time decision support, contributing to smarter urban planning and more resilient, sustainable cities. Furthermore, its participatory and data-driven design fosters community engagement and informed decision-making. Acknowledgements The authors would like to acknowledge the support of CESAM (UIDP/50017/2020 + UIDB/50017/2020 + LA/P/0094/2020) and C2TN (UIDB/04349/2020). Special thanks to FCT/MCTES for the research contract awarded to Helder Relvas (10.54499/2021.00185.CEECIND/CP1659/CT0026) and ODESSA - https://doi.org/10.54499/2024.07293.IACDC.