Lab 8: Data Frame Functions

Author

Instructions 

library(tidyverse)
library(palmerpenguins)
library(RColorBrewer)

Accessing the Lab

Download the template Lab 8 Quarto file here: lab-8-student.qmd

Important

Be sure to save this in the Lab 8 folder, inside your Week 8 folder, inside your STAT 331 folder!

Function to Standardize Variables in a Data Frame

Last week, you wrote a rescale_01() function which would rescale the values of a numeric variable to be between 0 and 1. This function worked on vectors, so to use it we needed to pair the function with mutate() if we wanted to make changes to the data. This week, we are going to use this function as a helper function for a larger rescale_column() function that will rescale variables in a given data frame.

Question 1: Create a rescale_column() function that accepts two arguments:

  • a data frame
  • one or more variable names to rescale

The body of the function should call the original rescale_01() function you wrote previously and return the original data frame with those columns replaced by their rescaled versions.

Important

Your function call must look like one of these two options:

# Tidy (unquoted) variable names
rescale_column(df, c(height, weight))

# Quoted variable names
rescale_column(df, c("height", "weight"))

To achieve this, you function must use one of the rlang (tidy evaluation) options from class.

Test Your Function: Use your rescale_column() function to rescale the bill_length_mm, bill_depth_mm, flipper_length_mm, and body_mass_g columns of the penguins dataset. Note: You may need to change the function inputs if you chose to use a character vector.

rescale_column(penguins, 
               cols = c(bill_length_mm, bill_depth_mm, flipper_length_mm, body_mass_g)
               )

Function to Create a Table

We’ve started to get more sophisticated with making tables, so let’s see if we can take these skills and create a function that outputs a pivoted table of counts.

Question 2: Write a pivot_table() function which takes in two categorical variables, counts the number of observations in each group, and outputs a table of counts that has been pivoted wider.

Your function should accept three arguments:

  • a data frame
  • a row variable (categorical variable to appear in the rows)
  • a column variable (categorical variable to appear in the columns)

Your function should:

  • count the number of observations for each combination of the two variables
  • pivot the results wider so that the row variable forms rows and the column variable forms columns
  • replace any missing counts (NA) with 0
  • add row and column totals using janitor::adorn_totals()
Important

Your function call must look like one of these two options:

# Tidy (unquoted) variable names
pivot_table(df, species, island)

# Quoted variable names
pivot_table(df, "species", "island")

To achieve this, you function must use one of the rlang (tidy evaluation) options from class.

Test Your Function

Use your pivot_table() function to display a table of counts for the species and sex of the penguins from the penguins dataset. Note: You may need to modify the names of the arguments!

pivot_table(penguins, row = sex, col = species)

Function to Create a Plot

In the chapter you saw a few different types of plotting functions—creating histograms with a specified binwidth, checking the linearity of a scatterplot, using hexagons to avoid over plotting in a scatterplot, and sorting bars from least to greatest in a barplot.

For this section, you are going to use these functions as inspiration to write your own plotting function.

Question 3: Your plotting function must meet the following criteria:

  • a data frame must be the first argument
  • two additional arguments for variables (can be numeric or categorical)
  • the plot must map one of the variables to color (or fill)
  • axis labels and the plot title should be descriptive (e.g., use labs() to replace raw variable names with human-readable labels)
  • return a ggplot object

You may choose the plot type (e.g., scatterplot, barplot, boxplot), as long as it meets the criteria above. You can also choose whether your function accepts quoted or unquoted variable names, but you must use one of the rlang (tidy evaluation) options from class.

Test Your Function

Test your function using variables from the penguins dataset.

Function to Style a Plot

For this final section, I want you to write a wrapper function that can style any ggplot object you create.

Question 4: Write a style_plot() function which accepts four arguments:

  • a ggplot object
  • a theme function (e.g., theme_bw())
    • the function should set a default theme (e.g., theme_minimal())
  • a vector of colors (e.g., RColorBrewer::brewer.pal(8, "Accent"))
    • the function should set a default color palette
  • a string indicating which aesthetic the colors should be applied to (“fill” or “color”)
  • a logical value indicating whether the legend should be included

Your function should:

  • Apply the given (or default) theme to the plot
  • Apply the colors to the specified aesthetic (fill or color)
  • Remove the legend (if requested to be removed)
  • Return the styled plot

Test Your Function

Carry out the two tests below for your style_plot() function. Note: You may need to modify the names of the arguments!

Test 1: My plot

I’ve created a barplot that is stored as my_plot. Test your function using this plot.

my_plot <- ggplot(penguins, mapping = aes(x = species, fill = sex)) + 
  geom_bar()

## This will use your default theme and colors
style_plot(my_plot, 
           aesthetic = "fill")

## This will use my favorite theme and colors
style_plot(my_plot, 
           theme_fn = theme_bw,
           palette = brewer.pal(3, "Paired"),
           aesthetic = "fill")

Test 2: Your plot

Copy the code from your test of your plotting function (@plot-function-test) and use your style_plot() function to style the plot! You are expected to pipe your plot into the style_plot() function. Do not use nested functions or save any objects!