Evaluating R Packages
Back in February I completed the Making Data Science Work for Clinical Reporting course from Genentech on Coursera, and I highly recommend it to any clinical programmers. As a member of the {riskmetric} development team, I found the course to be well-organized and up-to-date, with a nice emphasis on assessing and managing risk.
One topic that particularly caught my attention was the importance of evaluating R packages. Although it’s critical to use high-quality code in regulated industries such as pharma, the significance of risk assessment of R packages is often overlooked. To address this, the course introduced a helpful checklist for evaluating R packages:
- Does the package provide the required functionality?
- Is the package reliable?
- Does the code appear robust, and is it well-tested?
- Is the package well-documented?
- Who are the package authors, and are they responsive?
By answering such questions, we can quickly evaluate the quality and reliability of an R package. The course included a practice exercise that required us to assess an R package of our choice. Overall, the simple checklist is a valuable tool for anyone working with R packages in regulated industries. By ensuring that the packages we use meet high standards for functionality, reliability, robustness, and documentation, we can minimize the risk of errors and ensure data integrity. I highly recommend checking out the course to learn more.
