Meeting summary
The meeting focused on reproducible analysis workflows and tools, with Vahid introducing Quarto as a powerful tool for literate programming and discussing best practices for creating reproducible documents. The group explored challenges in reproducibility across different fields, sharing experiences in chemistry and psychology while discussing the difficulties of achieving perfect replication in human-based research. The session concluded with plans to organize workshops on Git and GitHub usage, along with discussions about AI tools in research and strategies for teaching reproducibility to students.
The meeting focused on reproducible analysis workflows, emphasizing the crisis of workflow reproducibility and the five pillars that support computational reproducibility. Vahid introduced Quarto as a powerful tool for literate programming, highlighting its benefits over traditional workflows and its role in creating transparent, executable documents. He also discussed best practices for creating reproducible documents, such as organizing with a clear narrative structure, ensuring code readability, and separating code from outputs. The session concluded with an open invitation for participants to share their experiences with documentation practices in their respective fields.
The group discussed challenges in reproducibility across different fields, with Kyle and Kimberly sharing their experiences in chemistry and psychology respectively. They explored how computational reproducibility is easier to document but data preprocessing and decision-making steps present significant challenges, particularly in fields involving human subjects. The discussion highlighted that while complete reproducibility is ideal, achieving perfect replication of human-based research decisions is inherently difficult due to the subjective nature of data interpretation.
The meeting focused on reproducibility in research, with Kimberly discussing her data sharing practices and Vahid presenting on tools and strategies for reproducible research. Vahid introduced version control, workflow orchestration, environment management, and documentation as key pillars for reproducibility, highlighting tools like Git, SnakeMake, NextFlow, and electronic lab notebooks (ELNs). The group discussed the importance of teaching reproducibility to students and strategies for curriculum integration. Vahid concluded by outlining an implementation roadmap for reproducibility, emphasizing the use of tools like Quarto and Git for new projects.
Vahid presented on project structure and reproducible science, emphasizing the importance of templates, automation, and FAIR principles for workflows. Brad expressed interest in the topics but noted some discomfort with tools like GitHub. The group discussed scheduling challenges, with Kimberly suggesting a poll for the last week of February to accommodate different schedules. Vahid agreed to create a poll, and Kyle inquired about the final outcome of the project, which Vahid outlined as sharing resources and potentially developing workshops or courses for PhD students.