Blended learning short course improves data science outcomes in environmental health

Public and environmental health professionals need data science skills to support research and practice. The Data Science for Environmental Health (DaSEH) short course program combines a two-week online course followed by a three-day, in-person codeathon. The program was designed to teach basic data science concepts in R and give learners the opportunity for real-world practice. We examined how the DaSEH program influenced learners' self-perceived and realized skills improvement with data science tools through surveys at three timepoints. We also assessed open-ended feedback for themes. Learners' comfort and confidence improved significantly after the online course but was not significant after the codeathon. However, summative assessment scores showed significant realized improvements in their ability to clean data, follow reproducible workflows, and plan analyses. Feedback from learners highlighted the importance of peer interactions, especially through codeathon "stand-ups". We suggest our blended learning model, consisting of an online course and an in-person codeathon, is a viable and scalable approach for data science professional development associated with improvement in learning outcomes.