A Call to Action
Keynote at R/Pharma 2023
October 26, 2023
Any opinions expressed in this presentation are solely my own and not necessarily those of my employer (Hoffmann-La Roche Ltd).
Partly based on the manuscript “Improving Software Engineering in Biostatistics: Challenges and Opportunities” (arXiv) which is joint work with:
Its content started from a panel discussion on “Research Software Engineering for Clinical Biostatistics” which took place at the 43rd Annual Conference of the International Society for Clinical Biostatistics (ISCB) in Newcastle in August 2022.
crmPack
crm
stands for continual reassessment method (Bayesian design for dose escalation)JAGS
code to run Markov Chain Monte Carlo (MCMC)openstatsware
mmrm
and brms.mmrm
)Many great tools exist to help with reproducibility of statistical analyses (see CRAN task view), e.g.:
Sweave
(2002)knitr
and Rmarkdown
(2014)packrat
(2014)officer
(2017)usethis
(2017)renv
(2019)targets
(2020)quarto
(2021)Similarly lots of tools have been provided for package developers:
testthat
(2009)devtools
(2011)checkmate
(2014)lintr
(2014)spelling
(2017)styler
(2017)pkgdown
(2018)precommit
(2020)Templates can be a great start to centralize code for reoccurring use cases, e.g.:
rOpenSci
openstatsware
have run five workshops in 2023These slides are at
danielinteractive.github.io/rpharma-2023-keynote
Feel free to connect at
linkedin.com/in/danielsabanesbove
Why We Need to Improve Software Engineering in Biostatistics | License