Week 4 Activity

R Basics (3) - EDA and ggplot2

Introduction

This activity puts to practice the concepts introduced in lecture today, including functions like:

  • dim()
  • sapply()
  • is.na()
  • sum()
  • summarize()
  • across()
  • group_by()
  • count()
  • theme_*()
  • ggplot()
  • aes()
  • geom_point()
  • geom_bar()
  • geom_col()
  • geom_histogram()
  • geom_boxplot()
  • geom_violin()
  • geom_density()

This activity is intended to be finished during classtime. In the case that we run out of time, please complete it indepedently and submit it via BCourses by Friday at 11:59PM. To submit to BCourses,

  • Render your QMD (Render button in RStudio and Preview button in Positron)  

  • Your HTML file will be created and can be found in the same folder on your computer that the QMD is stored in.  

  • Submit both the HTML and QMD files to the ‘Week 4 Activity’ assignment in BCourses.

If any of these functions are unfamiliar to you, you can access the help documentation by typing ?function_name into your R console. Alternatively, you can review R basics in the following textbook.

1. Load the required packages.

library(tidyverse)
library(palmerpenguins)

Hint: if you have not yet installed one of these packages, you must first run install.packages("package_name") in your R console.

We will use the penguins dataset from the palmerpenguins package for most questions and mtcars at the end.

Part 1: Understanding the Data (EDA Basics)

2. What are the dimensions of penguins? How many rows? How many columns?

Hint: Use dim().

3. What are the data types of each column?

4. How many total missing values are in the dataset?

5. What proportion of missing values are in each column? Which variable has the most missingness?

Part 2: Single Variable Distributions

6. Create a bar plot of species.

  • Use geom_bar().

  • Does this geom require you to count manually?

Written answer here:

7. Create a bar plot of island.

  • Which island appears most common?

Written answer here:

8. Create a histogram of bill_length_mm.

  • Does the distribution look symmetric, skewed left, or skewed right?

Written answer here:

9. Create a density plot of bill_length_mm.

  • Use geom_density().

  • How does it compare to the histogram?

Written answer here:

10. Create a boxplot of body_mass_g.

  • Are there potential outliers?

Written answer here:

Part 3: Grouped Summaries + geom_col()

11. Count the number of penguins in each species using count(). Store the result.

12. Use geom_col() to recreate the species counts.

  • Why must we use geom_col() instead of geom_bar() here?

Written answer here:

13. Count the number of penguins by species and island.

  • Which combination is most common?

Written answer here:

Part 4: Two Variable Relationships

14. Create a scatterplot of:

  • bill_length_mm (x)
  • flipper_length_mm (y)
  • What kind of relationship do you see?

Written answer here:

15. Modify the scatterplot:

  • Add color = species.
  • Does separating by species clarify the relationship?

Written answer here:

16. Change the same scatterplot to: 

  • shape = species instead of color.
  • When might shape be preferable to color?

Written answer here:

17. Create a boxplot of: body_mass_g by species. 

  • Which species tends to be heaviest?

Written answer here:

Part 5: Reflection (Short Discussion)

18. What is the difference between geom_bar() and geom_col()?

Written answer here:

19. What is the unit of observation in:

  • penguins?

  • mtcars?

Written answer here:

20. Which plots felt more exploratory vs explanatory?

Written answer here:

21. Next level

Let’s practice making some graphs with ggplot2!

  1. Load the mpg dataset (which is inside tidyverse package).  
  1. Take a look at the data: what is easy column representing? What does each row represent? (Hint: if you are not sure, type ?mpg into your console to launch the help panel)  
  1. Create a scatter plot of model vs cty, colored by year.  
  1. Experiment with different geom_() functions to visualize the data  
  1. Customize your plots with themes and labels.