R Packages and Libraries — Extending R's Power

Learning Objectives

By the end of this tutorial, you will be able to:

• Install, load, and manage R packages
• Understand the CRAN, Bioconductor, and GitHub package ecosystems
• Use the tidyverse meta-package effectively
• Create your own R package with devtools
• Troubleshoot common package issues

What Are Packages?

An R package is a bundle of functions, data, and documentation that extends R's capabilities.

Installing Packages

# Install from CRAN
install.packages("ggplot2")

# Install multiple packages
install.packages(c("dplyr", "ggplot2", "tidyr", "readr"))

# Install from GitHub
install.packages("devtools")
devtools::install_github("tidyverse/ggplot2")

# Install from Bioconductor
install.packages("BiocManager")
BiocManager::install("GenomicRanges")

# Install development version
devtools::install_github("user/package", ref = "dev")

Loading Packages

# Load package
library(ggplot2)

# Load with namespace
ggplot2::ggplot()

# See what's loaded
search()
# [1] ".GlobalEnv"        "package:ggplot2"   "package:stats"
# [4] "package:graphics"  "package:grDevices" "package:utils"
# [7] "package:datasets"  "package:methods"   "package:base"

# Detach package
detach("package:ggplot2", unload = TRUE)

Essential Packages

Tidyverse Ecosystem

# Install entire tidyverse
install.packages("tidyverse")

# Load it
library(tidyverse)
# Attaching core packages: ggplot2, tibble, tidyr, readr, purrr, dplyr, stringr, forcats

Data Science

Statistics

Package Information

# Package details
packageDescription("ggplot2")

# Available functions
ls("package:ggplot2")

# Vignettes
vignette(package = "ggplot2")
vignette("ggplot2-specs")

# Changelog
news(package = "ggplot2")

# Dependencies
tools::package_dependencies("ggplot2")

# Update packages
update.packages()

Creating R Packages

Package Structure

mypackage/
├── DESCRIPTION        # Package metadata
├── NAMESPACE          # Exported functions
├── R/                 # R code files
│   ├── function1.R
│   └── function2.R
├── man/               # Documentation
│   ├── function1.Rd
│   └── function2.Rd
├── data/              # Data files
├── tests/             # Test files
│   └── testthat/
└── vignettes/         # Vignettes

Using devtools

# Install devtools
install.packages("devtools")

# Create package
devtools::create("mypackage")

# Add function
devtools::use_r("function1")

# Add test
devtools::use_testthat()
devtools::use_test("function1")

# Add documentation
devtools::document()

# Build and check
devtools::build()
devtools::check()

# Install locally
devtools::install()

# Run tests
devtools::test()

DESCRIPTION File

Package: mypackage
Title: My Awesome Package
Version: 0.1.0
Authors@R: person("John", "Doe", email = "john@example.com", role = c("aut", "cre"))
Description: Does awesome things with R.
License: MIT + file LICENSE
Encoding: UTF-8
Roxygen: list(markdown = TRUE)
RoxygenNote: 7.2.3
Depends: R (>= 3.5.0)
Imports:
    dplyr,
    ggplot2
Suggests:
    testthat (>= 3.0.0),
    knitr,
    rmarkdown

Troubleshooting

Common Issues

# Package not found
install.packages("package_name")

# Version conflicts
packageVersion("ggplot2")

# Reinstall
remove.packages("ggplot2")
install.packages("ggplot2")

# Load conflicts
conflicts(detail = TRUE)

# Check conflicts
conflicts(filter = c("filter", "lag"))  # dplyr vs stats

# Use namespace to resolve
dplyr::filter()
stats::filter()

Package Dependencies

# See dependencies
tools::package_dependencies("tidyverse")

# Install all dependencies
devtools::install_deps("mypackage")

# Check for missing packages
missing <- setdiff(
  tools::package_dependencies("tidyverse")$tidyverse,
  installed.packages()[, "Package"]
)

Practice Exercises

Exercise 1: Package Explorer

Write a function that lists all functions in a package.

Solution

list_functions <- function(package_name) {
  if (!requireNamespace(package_name, quietly = TRUE)) {
    stop(paste("Package", package_name, "is not installed"))
  }
  ls(paste0("package:", package_name))
}

list_functions("ggplot2")

Key Takeaways

• install.packages() installs from CRAN — one-time per R version
• library() loads a package — every session
• Tidyverse is a meta-package — loads core data science packages
• Use devtools to create packages — modern package development workflow
• DESCRIPTION defines package metadata — name, version, dependencies
• devtools::document() generates documentation from roxygen2 comments
• devtools::check() runs comprehensive checks before submission

Next: Learn about R Data Import/Export — working with various data formats.