Week 6 Live Session (Fall 2026)

In this week 6 session we covered tidy data: the three rules, how to spot untidy data, and how to fix it with {tidyr} functions. David also coded live with a fake soccer training survey, splitting it into smaller data frames, analyzing them, and joining them back together, and we previewed the rest of the course.

You'll learn:

• How to check whether data is tidy by asking whether you could plot it, and by identifying your unit of analysis
• How to use pivot_longer() to turn year columns into year and value variables, and pivot_wider() to reshape results for a display table
• How to split multi-value cells with separate_wider_delim() and separate_longer_delim()
• How to break a messy survey data set into smaller tidy data frames that share a respondent ID, then rejoin them with left_join()
• How to turn non-numeric responses like 'skipped' into NA with parse_number() or when reading in data, and trim stray spaces with str_trim()

Resources mentioned:

• {tidyr}
• {dplyr}
• {ggplot2}
• Quarto
• Positron
• {stringr}
• {readr}
• {gapminder}
• {naniar}
• R for Data Science
• Claude Code
• Mermaid

Key moments:

• 02:11 Housekeeping and review of last week's AI session
• 04:07 Overview of the remaining course weeks
• 09:40 Participants share first reactions to tidy data
• 12:44 AI tools now default to tidyverse-style R code
• 18:30 Rule one: year spread across columns, fixed with pivot_longer()
• 21:15 Participant asks why pivot long and when to keep data wide
• 32:41 Splitting an address cell with separate_wider_delim()
• 34:07 Select-all-that-apply answers and separate_longer_delim()
• 39:26 Live coding: cleaning non-numeric survey responses to NA
• 41:49 Using parse_number() as an alternative to fix numeric columns
• 47:49 Participant question on whitespace, str_trim() and str_squish()
• 54:11 Joining demographics back in with left_join() to group by city
• 57:30 Participant asks how to handle missing data, {naniar} package