But here are some steps to try:

Step 1: Define the Problem

Think about what you were trying to do vs. what happened instead. Create a hypothesis for what went wrong.

Step 2: Read the Error Message.

Look for a line number where the error is occurring.

Sometimes the bug is in the line before the one that threw the error!

Step 3: Re-run code from the beginning.

Sometimes you ran something out of order. Go back to the beginning of your code and re-run to see if that will fix it.

Step 4: Talk it out!

Try your best to explain the problem out loud, preferably to a friend or teammate.

Practice rubber duck debugging.

Step 5: Check instructions, documentation, and Google.

If you’re lost, refer to all the resources you have: cheatsheets, lab guides, online documentation.

And when in doubt: Google the error message and see if someone else had the same problem!

Step 6: Ask for help!

Don’t let a single bug frustrate you for too long. If none of the above strategies worked, ask a classmate, TA, or instructor for help with the problem.

Save and/or Knit Often.

Remember, your RMarkdown won’t knit if there are bugs. This is a great way to find them early.

Use good names.

Name your variables and dataframes with care. Rename things to make them more clear. Good names can help you find a problem quickly.

Start simple, and build up little by little.

Don’t try to write a whole program all in one go.

Run your code line-by-line.

Check that it works as you go.

Leave yourself good annotations and comments!

Use # to leave comments: remind yourself what certain lines of code are doing.

Resources for Avoiding Errors

Practice Defensive Programming

Follow Style Guides:

Wickham’s Guide

Google’s Guide

Example 1

# Get just three columns of mpg.

small_mpg <- mpg %>%
  filter(cyll, trans, cty)

Example 2

# Make a summary table of highway fuel efficiency
# in different manufacturers.

group_by(mpg) %>%
  summarize(hwy)

Example 3

# Make boxplot of city fuel efficiency
# in different drive trains.

ggplot(aes(cty)) +
  geom_boxplot(aes(fill=blue))