911³Ô¹Ï

Using AI tools for R Coding

AI tools (such as ChatGPT) can be an excellent companion while learning R, but only if you know how to prompt it effectively. The guide below shows you how to use AI tools to write cleaner code, troubleshoot errors, and accelerate your learning while still understanding what you are doing.

NOTE: Generative AI software can be helpful in supporting coding activities. Before you begin using AI tools, we strongly recommend you familiarize yourself with any policies governing the use of these tools in your class, work/practicum placement, or with your community partners. 

1. Start with a Clear Research Task

AI tools work best when it knows exactly what you’re trying to do.

Good prompts include:

• Your research question
• The dataset you are using
• The variables (including names & types)
• The analysis you intend to run

EXAMPLE PROMPT

I’m analyzing survey data. My research question is: Are women more likely than men to report moderate-to-high loneliness?

Variables:

- gender  (factor: male, female, nonbinary)
- lonely_score (numeric 0–10 scale)
- lonely_cat (factor: low, moderate, high)

Please generate tidyverse code to calculate mean loneliness by gender and run a chi-square test on lonely_cat by gender. Also tell me if this is the correct test, or if an alternative analytic approach should be used.

2. Always Provide Variable Names, Types, and Sample Values

R code depends heavily on variable names and structures. Without them, AI has to guess and often guesses wrong.

EXAMPLE PROMPT

Here are the variables in my dataset:

- age (numeric)
- income (numeric, continuous)
- education (factor with HighSchool, College, University)
- health_status (factor, ordered: poor < fair < good < very_good < excellent)

Please write R code using tidyverse to:

1. Recode age into categories,
2. Summarize income by education, and
3. Plot health_status by education.

This ensures code is tailored to your dataset

3. Share Your Data Structure

Use str(df), glimpse(df), or a sample of your data.

EXAMPLE PROMPT

Here is glimpse(df) output. Use this structure to write code that…

Rows: 835
Columns: 6
$ id                         <int>
$ gender              <chr>
$ age                      <dbl>
$ lonely_score  <dbl>
$ lonely_cat       <chr>
$ province         <chr>

This helps your AI tool generate correct, runnable code.

4. Ask for Explanations, Not Just Code

You learn faster if your AI tool explains why something works.

GOOD EXAMPLE PROMPT

Write code to run a logistic regression predicting lonely_cat from age and gender.
Also explain what each line of code does, and how to interpret the coefficients.

5. Troubleshoot Errors With Copy/Paste

When R throws an error:

1. Copy the error message
2. Copy the code block you ran
3. Tell your AI tool what you expected to happen

ERROR TROUBLESHOOTING PROMPT

I ran this code:

df %>% mutate(lonely_score = as.numeric(lonely_score))

But I got the error:

Error: Problem with mutate(): NAs introduced by coercion

Can you explain why this happened and how I can fix it? Here is a sample of the data in lonely_score.

This is the single most useful pattern for debugging.

Sample "Perfect Prompt" Template

My goal: [state your research question]
Dataset: [name in R (e.g., df or data) + brief description]
Variables: [include name + type + sample values]
What I need: [graph? model? Cleaning?; explain in plain language]
Context: [assignment? thesis? practicum?]
Important constraints: [tidyverse only? no loops? Etc.]
Example Data Structure: [str() output or similar.]

Please provide code, an explanation, and an example output.