The Rebecca J. Holz Series in Research Data Management commemorates Research Data Services co-founder Rebecca Holz, who passed away unexpectedly in 2011.
Each talk will be held on a Wednesday from noon-1pm in Memorial Library 126. We invite you to bring your lunch!
Like to talk about your data? Have a topic you’d like us to present on? Please contact the RDS Outreach Committee.
To view previous presentations in the Holz series, check out our archive.
February 15 – Pierce Edmiston, PhD candidate, Department of Psychology, University of Wisconsin-Madison
Adopting open source practices for better science
What can scientists learn from the open source community when it comes to improving the reproducibility of research methods and results? I’ll introduce three problems in reproducibility, explain how they’ve been solved by the open source community, and demonstrate how these solutions can be utilized by scientists to make for better and more reproducible research. I’ll cover version control systems, dynamic documents, and “hollow research projects”–what I consider to be the epitome of reproducible research. I’ll review the empirical evidence that open source practices make for better science, and offer some speculation from the perspective of cultural evolution as to why open source has been so successful.
March 7 – Ankur R Desai, Professor, Atmospheric and Oceanic Sciences, Ned P Smith Professorship of Climatology, University of Wisconsin-Madison
Social coding: It’s not a communicable disease
We all work broadly among networks of scholars diffused among multiple institutions. Over time this has led to practices for sharing of ideas, data, and tools for group organization, meeting, and writing. At the same time, virtually all data analyses in all fields have become more computational and programming intensive. However, broadly accessible tools for joint coding and reproducing computational results of others have not been widely adopted by researchers. Social coding and development tools, such as GitHub, Docker, and Slack, have had widespread adoption in commercial and non-profit software engineering. Here, I will discuss how these tools also enhance research by improving interaction on large programming-based data analyses, ensuring code reliability and reproducibility, and providing a means to widely document and share code and results. I will draw on examples from my own lab’s research on development of community code for processing of environmental and atmospheric field observations and on the development of an informatics tool for ecosystem modeling. I will also highlight a colleague’s novel experiment to run an entire lab-based experiment from idea to instrument output to code to results to manuscript in GitHub. Finally, I will wrap up with discussion on the pros and cons of social coding in a research environment and what the future may look like for best practices in code development and sharing for research
April 26 – Isabelle Girard
Title and abstract forthcoming.