August 8-9, 2024 | In-person training

The Exposure Modeling Bootcamp is a two-day workshop focused on skills development in the application of both traditional and machine learning methods in predicting spatial/temporal variations in environmental exposures (e.g., air pollution, temperature, noise) using real data sets.

This two-day workshop is focused on practical skills development in modeling environmental exposures using both traditional and machine learning methods. The workshop is led by Dr. Scott Weichenthal (Associate Professor, McGill University) who has extensive experience in the development and application of exposure models in environmental epidemiology. Morning sessions will include lectures discussing important concepts related to exposure science and exposure modeling in environmental epidemiology and afternoon sessions will focus on hands-on laboratory exercises applying both traditional (e.g., linear regression, generalized additive models) and machine learning methods (e.g., random forest, neural networks) in modeling environmental exposures using real data sets. Participants will learn practical skills in working with environmental exposure data and will gain knowledge in the application of multiple approaches to modeling environmental exposures known to impact human health.

By the end of the workshop, participants will be familiar with the following topics:

• Principles of exposure science as applied to environmental epidemiology
• The intuition behind how various modeling approaches work including linear regression models, generalized additive models, random forest models, dense neural networks, and convolutional neural networks
• Data handling and cleaning
• Developing and evaluating predictive models
• Data collection and management for exposure models based on non-traditional data streams including images and audio data

Audience and Requirements

Investigators at all career stages are welcome to attend but we particularly encourage trainees and early-stage investigators to participate. There are four requirements to attend this training:

1. Each participant should have an introductory background in statistics (i.e., linear and logistic regression).
2. Each participant should be familiar with R/RStudio. All code examples used in the laboratory exercises will be annotated in detail but students will benefit from previous experience using R.
3. Familiarity with Python is an asset but is not required. We will use Python code in training convolutional neural networks but examples will be annotated in detail so students will understand what is happening without having to reproduce code on their own.
4. Each participant is required to have a personal laptop and a free, basic Posit Cloud (formerly RStudio Cloud) account. All lab sessions on the first day will be done using Posit Cloud (formerly RStudio Cloud).
5. Some lab sessions will use Google CoLab, so each participant is required to have a Google CoLab account (you will need a Google account to access Google CoLab).

Additional Information

• Subscribe for updates on new Boot Camp details and registration deadlines.
• Contact the Boot Camp team.

Capacity is limited. Paid registration is required to attend.