install.packages("psych")5 Installing and loading an R package
This week we aim to supplement the assigned readings and homework with a few blurbs and examples in R. We will learn how to install packages and load data into the RStudio environment, and will discuss how to build your own functions in R.
Two of the first skills that you are going to need to learn are 1) how to install / load your required packages into the R environment and 2) how to import data from a file. More often than not, these will be the first two steps of your analysis (though if you are working on more complex projects you may find it easier to keep track of things if you load import your data and packages as needed). In this walkthrough we will cover the basics of installing and loading packages. Importing data will be covered in the next.
5.1 There’s a package for that!
One of the great things about R is its extensibility via packages. For other great things about R (or how to sell R to your peers) see this link on simplystatistics.org. In fact this blog is a great resource for staying in contact with topics and new developments in data analysis.
From Peng’s article:
“With over 10,000 packages on CRAN alone, there’s pretty much a package to do anything. More importantly, the people contributing those packages and the greater
Rcommunity have expanded tremendously over time, bringing in new users and pushingRto be useful in more applications. Every year now there are probably hundreds if not thousands of meetups, conferences, seminars, and workshops all around the world, all related to R.”
For example, in my own work I often find myself not only using statistical techniques such as growth curve modeling (package: lmer) but also advanced quantification techniques such as cross recurrence quantificantion analysis (package: crqa), MultiFractal Detrended Fluctuation Analysis (package: MFDFA), and Sample Entropy (package: TSEntropies). Ten years ago this involved 100’s of lines of custom programming to carry out these analysis… today, there’s an R package for that—i.e., someone else likely with far greater programming abilities than I has already done it.
5.2 Installing a package using the GUI
For the point-and-click crowd, the easiest way to install an R package is to locate the Packages tab in your Window. From here you can click install and type in the package that you wish to install. So for example, let’s install the psych package, which is useful for obtaining descriptive stats from your data (although to be honest I hardly use it).
As you see the video, I left the Install dependences box ticked. You should always install with dependencies, this automatically installs other packages that may be required for psych (or whatever package of interest) to work. You may have also noticed when after you clicked OK, R entered this line of code into your Console.
Which leads us to…
5.3 Installing packages with install.packages()
A basic install of R comes with the base package which has an unbelievably large number of functions and analyses built in. However learning base R has a somewhat steep learning curve, especially for those with no programming experience. In this course we will use many of the simple functions in base R, but for more complex data wrangling (sorting, structing our data), statistical analysis, and plotting we will use the wonderful tidyverse package. tidyverse piggybacks on the basic install, replacing the sometimes bewildering R syntax with more natural language. Here, we’ll use tidyverse as an example of how to install from command line.
Installing tidyverse (or any package for that manner) couldn’t be easier. At the prompt (or in your notebook) simply type:
install.packages("tidyverse")Congratulations you have installed over 70 new packages to your R environment!
You see, tidyverse is not a single package (as is usually the case), but a collection of packages that function seamlessless with one another abiding a shared ethos of data science practices (who knew there were ethoses in stats! A simple google search of tidyverse might lead you to believe that you have joined a cult with Hadley Wickham as your leader!). The core value is that data should be structured, analyzed, presented, and shared in a manner that is as transparent as possible. In this class, we aren’t going to go the “full Hadley” (all the way down the rabbit hole) but we are going to abide many shared principles (it’s just good science!).
Note for the crqa package I simply type install.packages("crqa") and for MFDFA install.packages("MFDFA"). You noticing a pattern here?
5.4 An IMPORTANT note on Rendering with install.packages:
One quirk with knitting from your source is that you will need to specify which CRAN mirror you will download the package from. The Comprehensive R Network (CRAN) is an online repository that houses almost all things R, including packages. There are multiple mirrors set up all over the globe (https://cran.r-project.org/mirrors.html) and you need to tell R which one is your preferred choice. The quirk is that you only need to do this when knitting/compiling. You don’t need to do this for other everday use, but your source often will not compile unless you fix this. See here to read more about the error this produces.
Since your homework is going to involve rendering you should probably get in the practice of fixing this. There are two ways to resolve this issue. The first is to simply add the repos argument to your install.packages() command like so:
install.packages(package name, repos="http://cran.us.r-project.org")Executing this once (adding the repo) usually fixes things for the rest of your session. If you find you keep encountering this issue, there is a brute-force alternative that involves you providing a default repo at the outset. This can be accomplished by creating a new chunk at the beginning of your Quarto file (right underneath your header), and inserting the following:
setRepositories(graphics = getOption("menu.graphics"),
ind = NULL, addURLs = character())
r <- getOption("repos")
r["CRAN"] <- "http://cran.cnr.berkeley.edu/"
options(repos = r)5.5 Loading packages with library()
Once you have installed a package, it remains stored on your hard-drive until you delete it (which you’ll amost never do, packages take up so little disk space it’s really not efficient to install and uninstall unless absolutely necessary). What many new users have difficulty getting used to is that just because a package is installed does not mean that it is loaded. To load a package we typically use the library() command. For example to load tidyverse, we can type:
library(tidyverse)── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr 1.1.3 ✔ readr 2.1.4
✔ forcats 1.0.0 ✔ stringr 1.5.0
✔ ggplot2 3.4.3 ✔ tibble 3.2.1
✔ lubridate 1.9.2 ✔ tidyr 1.3.0
✔ purrr 1.0.2
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag() masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorsThat’s it, easy-peasy. Note that by default, only the base packages are typically loaded when you start a new session. Therefore, a useful practice to get into is to load the necessary packages at the start of every session.
Before moving on, I’d like to point out two things. First, notice that when installing with install.packages() we included quotations around the package name, but with library() we did not. Knowing when to use quotes and when not to is one of the more frustrating things to grasp for beginners (and even is an annoyance for me from time to time). All I can say is with practice, it becomes automatic. In truth, in the case of library() whether you include the quotes or not doesn’t matter (the default is to not), but for many functions, including install.packages(), the proper quotes do matter (for example: install.packages(tidyverse) will result in an error).
Second, while installing tidyverse installs over 70 packages on your computer, library(tidyverse) only loads 8 or so primary packages. This isn’t really much of an issue as we will be mostly using this common tidyverse packages, but it’s worth mentioning for future reference. Finally, when you loaded tidyverse it likely gave you a message about Conflicts. Conflict arise when different packages use the same names for functions. For example, the message:
`dplyr::filter() masks stats::filter()
is telling you that both stats and dplyr have a function called filter() and that the dplyr function is being used by default. This means that if you execute filter() by itself it will default to the dplyr version. If you ever find yourself in a situation where a function that you know should be working, isn’t working, first check to be sure R is indeed calling from the correct package. That said this can be avoided by getting in the habit of using the library::function() convention. I will typically use this format with a few exceptions (out of habit). For example, I will typically use ggplot() instead of ggplot2::ggplot().