osfr: An R Interface to the Open Science Framework

Abstract

Reproducible research requires effective project management workflows that promote consistency and facilitate transparency. Hallmarks of effective workflows include strategies for tracking the provenance of results, recording intermediate changes, conveniently documenting procedures, and working with collaborators without duplicating effort (Sandve, Nekrutenko, Taylor, & Hovig, 2013). For technically skilled researchers, the combination of version control software (VCS) such as git and cloud-based project repositories (e.g., GitHub and GitLab) enable highly effective workflows (Ram, 2013) that facilitate automation and computational reproducibility. However, these tools have a steep learning curve, especially for researchers whose training is far removed from software development. Alternatively, the Open Science Framework (OSF) offers much of the same functionality through an intuitive point-and-click web-based interface, significantly lowering the barrier to adopting best practices for researchers of all skill levels, or research groups composed of individuals with different levels of computational expertise (Sullivan, DeHaven, & Mellor, 2019). Yet the increase in accessibility comes at the cost of limiting opportunities for automation. osfr fills this gap for R users by allowing them to programmatically interact with OSF through a suite of functions for managing their projects and files.