Instructor: Alex Quistberg, Associate Research Professor; Drexel Dornsife School of Public Health
Dates: Monday, June 24
Time: 1:00 p.m. - 4:00 p.m. EST
Format: Online, live instruction
Built environment exposure data are key for epidemiological studies focused on walkability, road safety, physical activity, nutrition, violence, mental health and others, yet there many challenges to obtaining these data for researchers. Built environment data from geographic information systems (GIS) may be out of date, low quality or missing features and can vary substantially between the administrative entities that create them. Collecting built environment data manually or virtually via audits can be time consuming, expensive and may have limited geographic coverage. Artificial intelligence approaches via computer vision and deep learning offer researchers the potential to consistently collect built environment data at a much greater scale within and across geographies from publicly available street-level and satellite imagery. This workshop will provide attendees a broad and basic overview of the potential of AI for collecting built environment data from street-level imagery and provide examples of how they are being used for road safety research. The workshop will provide fundamentals about deep learning algorithms and models, how they compare to traditional data collection methods, what resources are needed to implement them, recommendations for collaboration with AI experts, finding and creating training data, and methodological aspects analysis and interpretation from the epidemiologic perspective. A brief tutorial will also be provided on how to set up, create and process training data for AI models.
As a result of this workshop, attendees will be able to:
- identify the best tools and costs for their built environment assessment needs;
- define and prioritize built environment items to collect using AI;
- prepare data for training AI models;
- describe how to conduct quality assessments of collected data; and
- how to find and collaborate with AI expert colleagues.
Prerequisite knowledge: None.
Technical requirements: None required, this course will cover conceptual, design and interpretation of artificial intelligence for public health research and not the programming and setting up a deep learning computing environment.
Continuing Education Credits*: 1.5 CEU or 15 CPH
