Assignment 7: R Functions and Packages

Author

Matthew DeHaven

Modified

March 12, 2025

1 Start a New Github Repository

You are going to make an R package that you are going to save and store on Github.

  • Make a new Github repository, with the name you want for your package
    • Make sure it’s public
    • You could name it “testPackage” if you would like1
  • Clone this new repository to your computer.

1 For R package names, avoid using special characters like “-” and “_” and spaces.

I am going to assume in these instructions that you named the package “testPackage”, but obviously if you named it something else, just replace the name.

2 Setup Your Package

  • Open up your local “testPackage” folder

  • Install “devtools”

  • In the terminal, run usethis::create_package(".")

This will create the package files in the current folder.

  • Add a new R script called “myfunc.r” to the “R/” folder

  • Add the following code to that script:

testPackage/R/myfunc.r
#' My first function
#'
#' @param a A numerical vector.
#' @param b Also a numerical vector.
#'
#' @return A numerical vector of a + b * a.
#' @export
#'
#' @examples
#' myfunc(3, 5)
myfunc <- function(a, b) {
  result <- a * b + a
  return(result)
}

  • In the terminal, run devtools::document()

  • In the terminal, run devtools::check()

    • Read the warning about the missing license

  • Add a license by running usethis::use_mit_license() in the terminal

  • Run devtools::check() again

  • Commit all the files, and push to Github

3 Install and Load Your Package

Now, we are going to see how we can use this package in other projetcs.

3.1 Accept Assignment on Github Classroom

  1. Accept Assignment 7 on Github Classroom.

  2. Clone this assignment to your computer.

  3. Initialize renv

3.2 Install Your Package

Since you made your package public on your Github account, you can install it using renv.

  • Install your package, renv::install("yourGithubUsername/testPackage")

  • Check to see if the following code works:

library(testPackage)
myfunc(3, 5)

  • Can you access the help file for your function?2

  • Check to see if you can see “testPackage” in the “renv/” folder.

2 If you are having trouble, go to the “R” pane in VSCode and click on “Clear Cache and Restart Helper Server”.

4 Assignment Tasks

4.1 New Functions

Estimating Beta

The task here is to write a new function to estimate \(\beta\) for a linear regression:

\[ y = \beta X + \epsilon \] \[ \Rightarrow \hspace{1cm} \widehat{\beta} = [X^\prime X]^{-1} X^\prime Y \]

Where \(y\) is a vector of \(n\) observations, \(X\) is a \(n\times k\) matrix of \(k\) variables, and \(\beta\) is a vector of \(k\) variables.

  • Write a function estimate_beta()
    • takes as inputs “y” and “X”
    • returns “beta_hat”

I am asking you to not use the lm() function in this task. You should write the code to calculate \(\widehat{\beta}\) yourself using linear algebra.

Custom Theme

Write a function that retuns your favorite ggplot2 theme.

You’ll need to use usethis::use_package("ggplot2") to add ggplot2 as a dependency. You may possibly need to add other packages if you want to use one of their themes as a starting point.

You should make at least 3 adjustments to a default theme.

  • Write a function my_theme()
    • returns your ggplot2 theme

Document, Check, and Push

  • Make sure to document your new functions

  • Run devtools::check() and see if you get any errors or warnings, try to fix them.

  • Push your package to Github

4.2 Using Your New Functions

Now you will use the functions you have written.

In your assignment folder,

  • Re-install your “testPackage” and load it.

Check estimate_beta()

  • Add the following code to your project, and run it to check your estimate_beta() function.
y <- iris[, "Sepal.Length"] |> as.matrix()
X <- iris[, c("Sepal.Width", "Petal.Length", "Petal.Width")] |> as.matrix()

estimate_beta(y, X)
                   [,1]
Sepal.Width   1.1210617
Petal.Length  0.9235289
Petal.Width  -0.8956758
  • Compare this to the official lm() coefficients
lm(Sepal.Length ~ 0 + Sepal.Width + Petal.Length + Petal.Width, data = iris)

Call:
lm(formula = Sepal.Length ~ 0 + Sepal.Width + Petal.Length + 
    Petal.Width, data = iris)

Coefficients:
 Sepal.Width  Petal.Length   Petal.Width  
      1.1211        0.9235       -0.8957

The “0” in the formula causes lm() to not include an intercept.

Check my_theme()

  • Add your theme to the following ggplot2 code
library(ggplot2)
p <-
  iris |>
  ggplot(aes(
    x = Sepal.Length,
    y = Sepal.Width
  )) +
  geom_point()
p

  • Save your plot to a PDF called “myplot.pdf”

5 Submit Your Assignment

  • Please add a link to your “testPackage” Github repository to the assignment’s README.md file.

  • Commit all your changes to the assignment repository and push it