Getting help

Nicky Wakim

Main ways to get unstuck

A huge part of coding is learning to debug your own code and figure out what’s going wrong.

The good news is that there are many ways to get help. These are some good steps (in order):

Use the ? in front of the function name to get more information!
- Best for: understanding function arguments and usage
Google or go to stackoverflow.com
- Best for: specific errors and edge cases
I can also go to my favorite AI tool to get help
- Best for: getting code started or explaining errors
Ask another person!
- Best for: when all else fails

1. R Documentation with `?`

Type ?function_name in the console to open the help page for any function.

?mean
help(mean)

1: These two commands do the same thing

Every help page has the same structure:

Section	What it tells you
Description	What the function does
Usage	The function signature
Arguments	What inputs it takes
Value	What it returns
Examples	Copy-pasteable code

Pro tip: Jump straight to Examples

Scroll to the bottom of any help page — the Examples section is often the fastest way to understand how a function works.

example(mean)

1: Run the examples directly

If ? doesn’t help, try searching across all installed packages:

??mean

1: Broader search of the function mean()

1. Demo: using `?` to solve a problem

Scenario: You want to round the number 3.14159 to 2 decimal places. You know there’s a round() function, but you’re not sure how to use it.

Step 1: Pull up the help page

?round

Step 2: Read the Usage and Arguments sections

Arguments and Examples are very helpful sections!

round(x, digits = 0)

x — the number to round
digits — how many decimal places (default is 0!)

Step 3: Try it out

round(3.14159)

1: Default: 0 decimal places

[1] 3

round(3.14159, digits = 2)

1: Now set digits to 2

[1] 3.14

2. Google and Stack Overflow

Tips for a good search:

Always include “in R” or “R programming” in your query
Copy and paste the exact error message into Google
Add the function or package name if relevant

Example: Getting NA from mean()

Google: "mean function returning NA in R"

This leads to answers like this Stack Overflow thread or this Posit Community post, where you’d learn about na.rm = TRUE.

x <- c(1, 2, NA, 4)
mean(x)
mean(x, na.rm = TRUE)

1: Returns NA:not what we wanted!
2: Returns 2.333...: ignores the missing value

[1] NA
[1] 2.333333

Reading Stack Overflow answers

Check the accepted answer (green checkmark) first
Look at the vote count: highly upvoted answers are usually reliable
Read the comments too: they often clarify edge cases
Check the date: R and its packages change over time
- Ex: mean( df$SleepDuration[df$Gender == 1] ) is a base R coding

2. Demo: Google and Stack Overflow

Scenario: You try to combine two pieces of text in R and get an unexpected error.

"Hello" + "World"

1: Seems reasonable… but + is only for math!

Error in `"Hello" + "World"`:
! non-numeric argument to binary operator

Step 1: Copy the error and search Google:

Error in “Hello” + “World” : non-numeric argument to binary operator in R

Googling the exact error message in quotes is often faster than describing the problem in your own words.

Step 2: Stack Overflow explains that + is only for math and Google’s search AI explains string combining uses paste() instead.

Step 3: Apply what you found

paste("Hello", "World")
paste0("Hello", "World")

1: Combines with a space in between
2: Combines with no space

[1] "Hello World"
[1] "HelloWorld"

3. AI Tools (ChatGPT, Claude, etc.)

When AI helps most:

Explaining what an error message means
Getting starter code for something new
Understanding what a function does
Adding intricate details to plotting
Helping with lengthy data management code

Example prompts:

“Why is the mean function in R giving me an NA?”

“Can you explain what na.rm = TRUE does in R?”

See an example ChatGPT conversation →

3. AI Tools (ChatGPT, Claude, etc.): some hesitation

If you do not want to use AI, that’s totally fine! You can still get help from the other methods

When using AI, watch out for:

AI often gives you more code than you need
AI can be confidently wrong: always test the code it gives you
AI may use functions we haven’t covered yet: stick to tools from class notes

Course policy on AI

If you cannot trace code back to the class notes, do not use it in assignments. You should be able to explain every line of your submitted code. See the syllabus for details.

3. Demo: Using AI effectively

Scenario: You want to find the largest value in a vector but have no idea what function to use.

A vague prompt (less useful):

“how do i get the biggest number”

AI may give a long answer with many options you don’t need yet.

A better prompt:

“In R, I have a vector of numbers: c(4, 7, 2, 9, 1). What is the simplest function to find the maximum value?”

AI will likely point you to max().

Step 2: Test it yourself — don’t just trust it!

scores <- c(4, 7, 2, 9, 1)
max(scores)

1: Always run the code to confirm it works

[1] 9

Step 3: Verify you understand it

?max

1: Check the docs to confirm what it does

AI gives you a starting point, not a final answer. Make sure you can explain every line before using it in your work.

4. Getting help from someone

When asking for help (from a classmate, from me, on Stack Overflow, or from AI), you want to include:
- Some information about the context
- A reproducible example (reprex)
These make it much easier for someone to help you!

Some information about the context

What you expected to happen
What actually happened
The exact error message you received

A good reprex has four parts:

Needed packages
Needed data
The exact code that produces the error
Needed code that leads to the error

In our labs, it’s good to save a .rds file of your qmd’s starting data, and share that with me

4. Demo: Writing a reprex

Scenario: You’re getting an unexpected result and want to ask your instructor or a classmate for help.

Focus on making your reprex self-contained and minimal: include what’s needed, but no more!

A bad help request:

“My mean function isn’t working, can you help?”

Nobody can debug this — there’s no code, no error, no data.

A good help request:

“I’m trying to calculate the mean of my numbers but keep getting NA. Here’s my code:”

library(tidyverse)
ages <- c(23, 31, NA, 45, 28)
mean(ages)

1: Packages included
2: Small, simple data included
3: The exact code producing the problem

[1] NA

The answer someone would give:

mean(ages, na.rm = TRUE)

1: na.rm = TRUE tells R to ignore missing values

[1] 31.75

Wrap-up

When you get stuck, try these steps in order:

Read the error message carefully — R is often telling you exactly what’s wrong
Check for typos — a missing comma, parenthesis, or quote causes most beginner errors
Use ? to check the function’s expected arguments
Google the error: paste it in with “in R” and see what comes up
Ask AI: describe your problem clearly; paste in your code and error
Ask a human: office hours, a classmate, or a forum with a reprex

If you come to me, I will ask you if you’ve tried all of the above first
- Not meant to be discouraging, but a huge part of coding is learning to debug on your own and determine when you need other human help

Even experienced programmers spend a lot of time debugging. Getting better at reading error messages is one of the most valuable skills you’ll build in this course.