Project Setup
project_setup.Rmd
In this section, we will quickly cover the environment, setup of the project, folder and file structure and supporting repositories/packages.
Environment
As we were a large team, it was decided for us all to use the same version of R packages needed for the creation of the TLFs and ADaMs. We also felt that this snapshot of packages used in our submission would more closely mimic a typical setup at a company delivering a R-based submission.
We made use of the renv R package to help us generate
a renv.lock
file to help developers configure their environments to be identical.
The renv.lock
has all primary and dependency packages versions that were installed in
our environments. We encourage readers to check out renv
package to find out more.
A synopsis of our environment:
- OS: Ubuntu 20.04
- R version: R 4.2.3
- Snapshot date: 2023-03-15
- Snapshot repository: https://packagemanager.posit.co/cran/2023-03-15
- Package Versions:
renv.lock
GitHub Repositories
We make use of several GitHub Repositories to store and collaborate for our work.
Core Analysis Repository
The website you are viewing is built on top of this Repository, which is where 95% percent of Pilot 3 work was done. The source code can be found on RConsortium/submissions-pilot3-adam. Below we discuss some of the key folders where our analysis was done as well as some files that we found useful for locking in our environments and keeping things reproducible.
-
adam/
: Original CDISC ADaM Pilot data created in SAS used for comparison -
submission/adam:
R generated ADaM datasets -
submission/output:
R generated TLFs -
submission/programs:
Production and QC programs for ADaMS -
submission/sdtm:
Original CDISC SDTM Pilot data created in SAS -
.Rprofile
: Project startup file to set up environment, -
renv.lock
: Locks in the versions of R packages and version of R -
vignettes/
: Markdown files that render TLFs and Datasets for Website
Utility Package Repository
Originally, the Pilot 3 Core Analysis Repository was a R package, but this quickly got messy and it was decided to split out the custom helper functions for the TLFs and Datasets to a separate R package.
The source code for our Utility R package can be found here.
Please note that this is not a CRAN package and is loaded in from
renv.lock
file via GitHub.