Practice Activity Week 4

Accessing the Practice Activity

Download the template Practice Activity Quarto file here: pa-4.qmd

Important

Be sure to save the file inside your Week 4 folder of your STAT 431 (or 541) folder!

Roundabouts

This week, our activity will explore how roundabouts are used around the world! You might not be as big of a fan of roundabouts as Dr. T, but these data allow us to investigate a fun and unique visualization technique—a Stankey diagram.

A Stankey diagram visualizes data flows. Specifically, the visual models the movement from one state to another. We are going to use a Stankey plot to visualize how traffic intersections are transformed into various types of roundabouts.

The roundabouts Package

The roundabouts R package provides a simple way to interact with the Roundabouts Database. You likely do not have the package installed on your computer, so the first step is to install the package:

pak::pak("emilhvitfeldt/roundabouts")

The roundabouts dataframe that is included in the package has an address variable which is a bit untidy. Below is some data cleaning code that will clean the address variable, and create three new variables: town_city, county_area, and state_region.

library(roundabouts)

roundabouts_clean <- roundabouts |>
  mutate(
    # Step 1: Remove strange <![CDATA[ tags from the beginning of the address field
    address = stringr::str_remove(address, stringr::fixed("<![CDATA[")), 
    # Step 2: Remove strang ]]> tags from the end of the address field
    address = stringr::str_remove(address, stringr::fixed("]]>")),
    # Step 3: Isolate the country at the end of the string between the ( ) symbols
    country = stringr::str_extract(address, "\\([^)]+\\)$"),
    # Step 4: Remove the brackets surrounding the country name
    country = stringr::str_remove_all(country, "[\\(\\)]"), 
    # Step 5: Remove the country from the address field
    address2 = stringr::str_remove(address, "\\s*\\([^)]+\\)$")) |>
  # Step 6: Separate address into town, county, state by comma
  separate(address2, 
                  into = c("town_city", "county_area", "state_region"), 
                  sep = ",") |> 
  # Step 7: Fix one instance of California instead of CA
  mutate(state_region = str_trim(state_region, side = "both"), 
         state_region = fct_recode(state_region, 
                                   "CA" = "California")
         ) |> 
  # Step 8: Remove companies from states
  filter(! str_detect(state_region, pattern = "Co.")) |> 
  # Step 9: Grab columns we will be using
  select(town_city, county_area, state_region, country, 
         year_completed, type, previous_control_type, control_type)

Creating a Region Variable

Alright, we now have the country of each roundabout, but that seems to fine grain for our analysis. If we’re interested in how roundabouts are used around the world, it seems like a region variable might be a better fit.

  1. Create a new continent column in the dataset, which classifies each country into one of the five major continents—Americas, Europe, Asia, Oceania, Africa.

The countrycode() function from the countrycode package will be helpful here! Here is an example of how you can use this function to obtain the continent of a given country:

countrycode::countrycode("Brazil",
            origin = "country.name",
            destination = "continent")
  1. You should have gotten a message about 1 matching error. Meaning, there was one specific country that couldn’t find the name for. Modify your code to manually classify the continent of this country.

Exploring Roundabouts

Let’s explore the data a bit before we jump into visualizing these data.

  1. What countries have the most roundabouts (in these data)? The fewest roundabouts?
  1. In these data, what types of roundabouts are most common?
  1. How does the distribution of types differ by continent?
  1. We’ve seen a lot of traffic calming circles showing up around SLO. What states in the US have the largest proportion of traffic calming circles?

The ggalluvial Package

Now that we have the data ready, let’s explore making Stankey plots with the ggalluvial R package.

We are interested in visualizing how roundabouts are incorporated into traffic intersections. To do this, we will visualize the flow from the previous intersection type to the type of roundabout that was added, to the final intersection (control) type.

Let’s start by creating the data gglluvial requires for plotting.

  1. Create a table of frequencies of previous intersection types, current intersection types, and roundabouts. Let’s focus on intersections (both old and new) where the type was known (not labeled as "Unknown").
  1. Now create the Stankey plot using the data you made above! Your axes should be previous control type and the current control type, and the colors should be associated with the type of roundabout. *Hint: You might find the
  1. Now that you have a working solution, let’s make the strata labels easy to read. Look over this reference article on different ways to label the strata for each axis. Implement your favorite method!
  1. Final touches! Let’s add some custom colors and an informative plot title! Each of the axes in your plot should have a title as well!