R Installation and User Libraries

R Installation and User Libraries#

R is the program that runs R code. R opens an interactive console; Rscript runs a saved script from a terminal. RStudio and Visual Studio Code are optional editors: neither installs nor replaces R itself.

If you are using a managed course computer, use its installed R first. Do not reinstall or upgrade managed software during a class unless teaching staff ask you to.

Check before installing#

Open a terminal and run:

R --version
Rscript --version

If both commands report a version, continue to Run a clean project check. If an editor cannot find R but these commands work, fix the editor configuration rather than installing a second copy.

Install R#

Use the official Comprehensive R Archive Network (CRAN) downloads.

macOS: use R for macOS and choose the installer that matches your Mac and supported macOS version. Open a new terminal afterwards.
Ubuntu: follow the CRAN Ubuntu instructions when you need CRAN’s current release. On a managed machine, ask the administrator. For a simple distribution-provided installation on a machine you administer, sudo apt update followed by sudo apt install r-base is sufficient, but the Ubuntu release may provide an older R version. See Linux Setup and Ubuntu Administration before changing repositories or system packages.

Install R before an editor. If you want RStudio Desktop, use the current Posit download page and check its supported platforms. For VS Code, follow VS Code for MQB after the terminal checks above work.

Install packages without running R as administrator#

R searches one or more package libraries. Inspect them in the R console:

.libPaths()
Sys.getenv("R_LIBS_USER")

Install course packages from an ordinary R session:

install.packages("package_name", repos = "https://cloud.r-project.org")

If R asks to create a personal library, answer yes. This is the normal, recommended location for packages that belong to your account. If needed, create R’s default user-library directory and restart R:

dir.create(Sys.getenv("R_LIBS_USER"), recursive = TRUE, showWarnings = FALSE)

Do not start R with sudo, and do not make a system library writable to everyone. Administrator access may be needed to install R itself or a missing operating-system development library, but it is not the normal way to install an R package. If an installation fails, read the error before adding compilers or system libraries; it should identify the missing requirement.

Run a clean project check#

Create a disposable folder with code, data, and results subfolders. Save this as code/r_setup_check.R:

cat(R.version.string, "\n")
cat("Working directory:", getwd(), "\n")
cat("First library:", .libPaths()[1], "\n")

dir.create("results", showWarnings = FALSE)
png("results/r_setup_check.png", width = 640, height = 480)
plot(cars, main = "R setup check")
dev.off()

stopifnot(file.exists("results/r_setup_check.png"))
cat("R setup check passed\n")

In a terminal, change to the project root, then run:

Rscript code/r_setup_check.R

The command should print R setup check passed and create results/r_setup_check.png. This is MQB’s default script convention: start at the project root unless an exercise explicitly names another directory. Paths such as data/input.csv and results/figure.png are then portable. Do not put a computer-specific setwd() in a shared script.

When something goes wrong#

When asking for help, copy the exact error and report your operating system plus the output of:

R.version.string
R.home()
.libPaths()
Sys.which(c("R", "Rscript"))

Also say whether the failure occurs in a terminal, RStudio, VS Code, or a notebook. An R notebook additionally needs the separate The R kernel; a Python environment does not install or isolate R packages.