Registration & Fees

As urbanization and climate change accelerate, accurate air quality monitoring has become increasingly important. Traditional monitoring systems rely on sparse, point-based ground stations, creating substantial spatial gaps in exposure assessment. This workshop provides a comprehensive guide to bridging those gaps by leveraging the cloud-computing power of Google Earth Engine (GEE) alongside advanced artificial intelligence. Participants will learn how to seamlessly integrate vast repositories of satellite imagery variables with cutting-edge Machine Learning (ML) algorithms—specifically Random Forest and Gradient Boosting—to predict air pollution concentrations.

The workshop will cover the full workflow of multicentre studies, from data collection and harmonization to exposure assessment, statistical modelling, evidence synthesis, and interpretation. It will also examine collaborative practices required for such studies, including common analytical protocols, shared data procedures, and reproducible code. Particular attention will be given to pooled analysis approaches and one-stage and two-stage designs, with emphasis on their assumptions, advantages, limitations, and practical implications.

The primary focus is introducing the concept and interdisciplinary methods of exposome science, which studies the totality of environmental exposures (physical, chemical, biological, lifestyle, and psychosocial) over an individual's life-course. The sessions cover exposome analytics, internal versus external exposomes, and practical applications in clinical epidemiology and birth cohorts.

This workshop will provide a practical, hands-on introduction to mixture methods commonly used in environmental health research, with emphasis on selecting appropriate methods, implementing analyses, interpreting findings, and communicating results clearly. The workshop will use applied examples from environmental epidemiology, including PFAS, pesticides, metals, air pollutants, and endocrine disrupting chemicals. Participants will learn how mixture methods can help answer key questions: What is the overall mixture effect? Which components contribute most strongly? How should results be interpreted when exposures are correlated? What are the limitations of different methods? The session will combine short lectures, guided demonstrations, hands-on exercises, and case-study discussion. Example datasets and annotated R code will be provided.