1 Towards a structured workflow
What is it actually?
Briefly put, the meaning of ‘workflow’ in a general sense may be summarized as
“[…] an abstraction of an automated and computerized business process. It consists of a set of activities that are interconnected by control flows […]. Each activity is a naturally defined task in a workflow and has associated servers that are either humans or executions of programs commonly called processes […].” (Son and Kim 2001)
Accordingly, from an R perspective, the term typically describes a set of interconnected scripts and functions which sequentially perform individual tasks to generate a desired result. In the majority of cases, such workflows rely on certain input data based on which some statistical metrics or figures shall be produced.
What’s the use of it?
This section will introduce you to two fundamental pillars of a structured workflow in R, namely
While the former is meant to separate the single projects you are working on and, at the same time, keep your stuff together in unique project-related working directories, the latter helps you to keep track of your single R code snippets and, more precisely, the changes you (or somebody else) made over time.
References
Son, Jin Hyun, and Myoung Ho Kim. 2001. “Improving the Performance of Time-Constrained Workflow Processing.” Journal of Systems and Software 58 (3): 211–19. doi:10.1016/S0164-1212(01)00039-5.