RStudio, without the panic

The big idea

R is the engine: a language that does maths, handles data and draws graphs. RStudio is the dashboard you drive it from. You never really "use R" on its own — you type R instructions into RStudio, and RStudio passes them to R.

Think of R as a very fast, very literal assistant. It does exactly what you write, in the order you write it, and nothing else. It can't guess what you meant. A missing comma, a capital letter in the wrong place or a file it can't see will stop it — and it tells you so with an error message. Errors are not you failing; they are R asking for clearer instructions.

The four panels

RStudio splits the screen into four areas. Once you know which one does what, it stops feeling cluttered.

Script — top left

Your recipe. This is where you write and keep instructions. It's a file (ending in .R) that you save, so you can come back next week and repeat everything. If the file name shows in red with an asterisk, it has unsaved changes.

Console — bottom left

Where R actually runs things and shows the results. Each instruction appears after a > prompt with the answer underneath. Think of it as R talking back to you.

Environment — top right

R's memory. Every object you create (a dataset you've loaded, a number you've stored) is listed here. If something isn't in this list, R doesn't know about it — which explains a lot of errors.

Files, Plots and Help — bottom right

A file browser for your space on the server (this is where you upload the weekly zip), the place graphs appear, and the built-in help pages. Type ?mean in the console and the help for mean() opens here.

Script, not console

This one habit prevents more confusion than anything else on this page.

To run a line from the script, click anywhere on it and press CtrlCmd + Enter (or click the Run button at the top of the script panel). R runs that line and moves the cursor to the next one, so you can step through a script by pressing it repeatedly — just like the demo at the top of this page.

Your weekly routine

Every PSYC121 lab follows roughly the same pattern. Once this is automatic, you can spend your brainpower on the statistics instead.

1. Get connected

   On campus, use Eduroam. Off campus, switch on the university VPN first — the server won't load without it.

2. Log in to the RStudio server

   Go to psy-rstudio.lancaster.ac.uk, choose Sign in with SAML and use your normal university login. Everything lives on the server, so you'll see the same files from any computer.

3. Make a folder for the week

   In the Files tab (bottom right), click New Folder, call it something like week_4, then click into it. One folder per week keeps things findable.

4. Upload the zip

   Download the week's zip from the lab page, then in Files click Upload and choose it. The server unpacks it for you, so you'll see the script and data files appear.

   On a Mac, Safari sometimes unzips downloads automatically. If you end up with a folder instead of a .zip, either upload the files inside it one by one, or turn off "Open 'safe' files after downloading" in Safari's settings.

5. Open the script and Save As

   Click the .R file to open it. Use File ▸ Save As to make your own copy (e.g. week_4_freya.R) so you keep an untouched original.

6. Point R at the right folder

   Choose Session ▸ Set Working Directory ▸ To Source File Location. This tells R "look for data files in the same folder as this script". Skipping this is the number one cause of "file does not exist" errors.

7. Load the tidyverse

   Run library(tidyverse) first, every session. It switches on the extra tools (read_csv, %>%, ggplot and friends). The wall of red-ish messages it prints is normal — it's not an error.

8. Work through the script

   Run line by line, read what comes out, and add your own # notes saying what each bit does in your own words. Those notes are revision gold before the class tests.

9. Save, then switch off

   Save the script (CtrlCmd + S), then click the red power button at the top right to end your session cleanly.

Reading R code

Almost every line you'll see in first year is built from the same handful of pieces. Learn to spot these and scripts start to read like sentences.

cows <- read_csv("penelope22.csv")

mean(cows$estimate, na.rm = TRUE)

# average guess, ignoring missing answers
mean(cows$estimate, na.rm = TRUE)

Pipes in plain English

Week 3 introduces chains like the one below. They look intimidating, but read top to bottom, saying "and then" at every %>%, they're just a list of instructions.

The result is a small table with one row per group. Swap mean for median or sd, or swap identity for another grouping column, and you've got a different analysis — that's what most lab exercises are really asking you to do.

Inside a pipe you don't need the cows$ part. Because the data has already been passed in, you can just write the column name.

A few more verbs you'll meet

# keep only some rows
cows %>% filter(estimate > 1000)

# add a new column calculated from another
cows %>% mutate(estimate_kg = estimate * 0.4536)

# count how many rows are in each group
cows %>% count(identity)

Graphs with ggplot

ggplot builds a graph in layers, like stacking transparent sheets. You always start with the data and say what goes on each axis, then add a layer that decides the shape.

cows %>%
  ggplot(aes(x = identity, y = estimate)) +   # the data and the axes
  geom_boxplot() +                            # draw boxplots
  labs(title = "Guesses of Penelope's weight",
       x = "Gender identity",
       y = "Estimate (lbs)")                  # tidy labels

aes() is short for "aesthetics" — which column maps to x, which to y, which to colour. Swap geom_boxplot() for geom_violin() or geom_jitter() and the same data gets drawn a different way. For a single column, try ggplot(aes(x = estimate)) + geom_histogram().

Shortcuts worth learning

You can do everything with menus, but these few save a lot of clicking.

Because RStudio runs in a web browser, a few shortcuts can be grabbed by the browser instead. If one doesn't work, use the menu or the Run button.

Check yourself

Six quick questions on the ideas above. Nothing is recorded.

When you're stuck

Getting stuck is part of learning to code, not a sign you're bad at it. Things to try, roughly in this order:

1. Read the red text slowly. It usually names the exact object, file or symbol it choked on. Check the error decoder.
2. Restart and re-run. Session ▸ Restart R, then run from line 1.
3. Check the help. ?summarise in the console opens the help page, with examples at the bottom.
4. Redo the learnr pre-lab tutorial for that week. They are built to practise exactly what the lab needs.
5. Ask a coursemate or a GTA in the lab. They see the same five errors all day.
6. Post on the PSYC121 Moodle forum or bring it to the weekly Q&A session. If you're stuck, dozens of others are too.

How to ask so you get a useful answer

Always include the code you ran and the exact error. That alone usually gets you an answer in one reply.

Going further

• PSYC121 lab materials — the department's own weekly lab sheets.
• R for Data Science — free online book that the course itself recommends; uses the same tidyverse style.
• Howell, Fundamental Statistics for the Behavioral Sciences — the department's endorsed textbook, with R sections. In the library.