Tidy Evaluation

Lecture 28

Dr. Eric Friedlander

College of Idaho
CSCI 2025 - Winter 2026

The Problem

Indirection

• Interactive R: We type diamonds |> filter(carat > 1).
  • carat is a data-variable (lives in the dataframe).
• Shiny R: We have input$var which contains the string "carat".
  • input$var is an env-variable (lives in the environment).
• Issue: filter(diamonds, input$var > 1) tries to compare the string "carat" to 1, which fails or does weird things.

Data Masking

The .data Pronoun

• Solution: Tell dplyr or ggplot2 exactly where to look.
• .data[[ string ]]: Look up the column named string in the data frame.
• .env$var: Look up the variable var in the environment (safer than implicit lookup).

Example: Filtering

server <- function(input, output, session) {
  
  filtered_data <- reactive({
    # OLD/WRONG: filter(df, input$var > input$min)
    
    # CORRECT:
    df |> filter(.data[[input$var]] > .env$input$min)
  })
  
}

Example: Plotting

• aes() also uses data masking.
• Use .data[[ ]] to map inputs to aesthetics.

output$plot <- renderPlot({
  ggplot(df, aes(.data[[input$x]], .data[[input$y]])) +
    geom_point()
})

Tidy Selection

Tidy Selectors

• Functions like select(), pivot_longer(), across() use tidy selection.
• Interactive: select(df, carat, price).
• Programmatic: select(df, all_of(input$cols)).
  • all_of(): Errors if any column is missing (strict).
  • any_of(): Ignores missing columns (lenient).

# If input$cols is c("carat", "price")
selected_data <- reactive({
  df |> select(all_of(input$cols))
})

Wrap Up

Rules of Thumb

1. Data Masking (filter, mutate, aes, arrange):
   • Use .data[[input$name]] to refer to a column by string.
2. Tidy Selection (select, pivot_, across):
   • Use all_of(input$names) to refer to columns by a vector of strings.

Do Next

1. Read Chapter 12: Tidy Evaluation from Mastering Shiny.
2. There’s NO recitation Gem for this textbook but I recommend creating your own and adding the textbook chapter and these slides.
3. I reommend working through the Exercises in the textbook.
4. Move on to lecture 29!