Containerization and R for Reproducibility and More

Abstract

Containerization has become a dominant computing paradigm for computing in the past decade due to its many advantages: isolation and security, scalability and efficiency with lightweight containers sharing an operating kernel and resources, and portability across cloud computing providers. For the researcher, analyst, or R user, containers have applications ranging from reproducible analytical environments to packaging statistical code to use in web applications. I will discuss how biomedical researchers can make use of containerization technology, particularly the tools provided by the Rocker Project, which publishes powerful standardized containers for the R language.

Speaker

Noam Ross is a computational disease ecologist and Executive Director of rOpenSci, a nonprofit dedicated to promoting open science and validating data science and computational methods. He is a core member of the Rocker Project, which maintains standardized containers for the R computer language. Noam’s work includes spearheading rOpenSci’s work in software peer review, developing a widely emulated system for leveraging the academic peer-review process coupled with state-of-the art automated code analysis to improve code quality in the scientific software in ecosystem, as well as using review as a mechanism for community building and training. His research interests and contributions span a wide range of topics, including disease ecology, zoonotic spillover, mechanistic modeling of disease dynamics, and non-parametric data science methods. His applied work includes creating early outbreak assessment models for the U.S. Defense Threat Reduction Agency, and modeling and forecasting for New York State’s COVID-19 emergency response. Noam holds a Ph.D. in theoretical ecology from the University of California-Davis and a B.Sc. from Brown University.