Creating an R Package

Math 241, Spring 2026

Step 1: Create a version control R package.

• There are a couple of ways to do this step. For this demo and Project 2, we are going to clone a starter repo from GitHub.
• In the future, you should consider using usethis::create_package().

Step 2: Create a version control RStudio Project with the same name as the GitHub repo.

• Should now have a Git pane and a Build pane in the upper-right window of your RStudio session.
  • Restart your R Session if either is missing.

Step 3: If you want to include data in the package, add the raw data and wrangle it into the version the package will share.

• Run usethis::use_data_raw() to create a data-raw folder and the DATASET.R file.
• Add the raw data files to the data-raw folder.
• Use the file DATASET.R to load and wrangle the raw data.
  • Instead of loading packages with library(package_name), use package_name::function_name().
  • At the bottom, include the following code to create a tidy .Rda file:

usethis::use_data(insert_data_name, overwrite = TRUE)

• Run the code in DATASET.R to create a new folder called data that contains the tidy data.
• Run usethis::use_r("insert_data_name") to create a blank script file.
  • We will add the data codebook to this file in Step 5.

Step 4: Add your functions to the R folder.

• For each user-facing function in your package, run usethis::use_r("insert_function_name") to create a new script file.
• Within each function, instead of loading packages with library(package_name), use package_name::function_name().
• For any packages your functions depends on, run usethis::use_package("package_name").

Step 5: Create documentation.

• For each of your function scripts in the R folder, add some template documentation code by going to Code > Insert roxygen skeleton.
• For each of your data scripts, you will need to write it from scratch.
  • Mimic examples!
  • Make sure to include @format and @source.
  • Include the dataset name in quotes at the bottom of the script.
• Add roxygen comments that document your function or dataset.
  • Here’s a data example.
  • Here’s a function example.
  • See the Object Documentation Chapter of R packages for more information on the syntax.
• To create the output help files, run devtools::document().
  • Notice that there is now a man folder with Rd help files.

Step 6: Test drive your package functions.

• Restart your R Session and run devtools::load_all() to make the package functions and data available.
• Test out the functions.
  • Return to earlier steps if you find any bugs.
• Type ?insert_function_name or ?insert_data_name to make sure the help file pops up and to see if it is formatted correctly.

Step 7: Run a more formal check of your package with devtools::check().

• Fix any errors or warnings. (Note: The package will still compile when there are warnings and notes.)

Step 8: Try installing the package with devtools::install().

Step Often: Commit and push the changes to the GitHub repository.

Steps to Come:

• Package metadata
  • DESCRIPTION
  • NAMESPACE
  • License
• Documentation
  • Best practices for the help files
  • Vignettes
  • Effective Readme
  • Website
• Testing
• Deleting or ignoring extraneous files