Getting started

• Go to the course GitHub organization and locate your HW 02 repo, which should be named hw-02-plastic-waste-[GITHUB USERNAME].

• Click on the green Code button, select the second option “Open with GitHub Desktop”

• The GitHub Desktop application will open up, with a white window that says “Clone a Repository”. Important: in the second line that says “Local Path”, there is a button that says Choose.... Click on it, and find and select the folder we created for this course. Then hit the blue Clone button.

• After successfully cloning, the window will disappear and you will see the that Current Repository is the one you just cloned. Success!

• Navigate to the project folder you just created within the Math118 folder, and open the hw-02-plastic-waste.Rmd file to begin.

Packages

We’ll use the tidyverse package for this analysis. You can run run the following code in the Console to load this package.

library(tidyverse)

The data

The dataset for this assignment can be found as a csv file in the data folder of your repository. You can read it in using the following line of code. Make sure the file path is pointing to this project folder! If you are not sure: hit “Session” -> “Set Working Directory” -> “To Source File Location”.

plastic_waste <- read_csv("data/plastic-waste.csv")

The variable descriptions are as follows:

• code: 3 Letter country code
• entity: Country name
• continent: Continent name
• year: Year
• gdp_per_cap: GDP per capita constant 2011 international $, rate
• plastic_waste_per_cap: Amount of plastic waste per capita in kg/day
• mismanaged_plastic_waste_per_cap: Amount of mismanaged plastic waste per capita in kg/day
• mismanaged_plastic_waste: Tonnes of mismanaged plastic waste
• coastal_pop: Number of individuals living on/near coast
• total_pop: Total population according to Gapminder

Exercises

Let’s start by taking a look at the distribution of plastic waste per capita in 2010.

ggplot(data = plastic_waste, aes(x = plastic_waste_per_cap)) +
  geom_histogram(binwidth = 0.2)

One country stands out as an unusual observation at the top of the distribution. One way of identifying this country is to filter the data for countries where plastic waste per capita is greater than 3.5 kg/person.

plastic_waste %>%
  filter(plastic_waste_per_cap > 3.5)

## # A tibble: 1 × 10
##   code  entity     conti…¹  year gdp_p…² plast…³ misma…⁴ misma…⁵ coast…⁶ total…⁷
##   <chr> <chr>      <chr>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
## 1 TTO   Trinidad … North …  2010  31261.     3.6    0.19   94066 1358433 1341465
## # … with abbreviated variable names ¹​continent, ²​gdp_per_cap,
## #   ³​plastic_waste_per_cap, ⁴​mismanaged_plastic_waste_per_cap,
## #   ⁵​mismanaged_plastic_waste, ⁶​coastal_pop, ⁷​total_pop

Did you expect this result? You might consider doing some research on Trinidad and Tobago to see why plastic waste per capita is so high there, or whether this is a data error.

1. Plot, using histograms, the distribution of plastic waste per capita faceted by continent. What can you say about how the continents compare to each other in terms of their plastic waste per capita?

⊕From this point onwards the plots / output of the code won’t be printed in the homework, but you can run the code and view the results yourself.

Another way of visualizing numerical data is using density plots.

ggplot(data = plastic_waste, aes(x = plastic_waste_per_cap)) +
  geom_density()

And compare distributions across continents by coloring density curves by continent.

ggplot(data = plastic_waste, 
       mapping = aes(x = plastic_waste_per_cap, 
                     color = continent)) +
  geom_density()

The resulting plot may be a little difficult to read, so let’s also fill the curves in with colors as well.

ggplot(data = plastic_waste, 
       mapping = aes(x = plastic_waste_per_cap, 
                     color = continent, 
                     fill = continent)) +
  geom_density()

The overlapping colors make it difficult to tell what’s happening with the distributions in continents plotted first, and hence covered by continents plotted over them. We can change the transparency level of the fill color to help with this. The alpha argument takes values between 0 and 1: 0 is completely transparent and 1 is completely opaque. There is no way to tell what value will work best, so it’s best to try a few.

ggplot(data = plastic_waste, 
       mapping = aes(x = plastic_waste_per_cap, 
                     color = continent, 
                     fill = continent)) +
  geom_density(alpha = 0.7)

This still doesn’t look great…

1. Recreate the density plots above using a different (lower) alpha level that works better for displaying the density curves for all continents.

2. Describe why we defined the color and fill of the curves by mapping aesthetics of the plot but we defined the alpha level as a characteristic of the plotting geom.

There is yet another way to visualize this relationship is using side-by-side box plots.

ggplot(data = plastic_waste, 
       mapping = aes(x = continent, 
                     y = plastic_waste_per_cap)) +
  geom_boxplot()

1. Learn something new: violin plots! Read about them at http://ggplot2.tidyverse.org/reference/geom_violin.html, and convert your side-by-side box plots from the previous task to violin plots. What do the violin plots reveal that box plots do not? What features are apparent in the box plots but not in the violin plots?

⊕Remember that we use geom_point() to make scatterplots.

1. Next, visualize the relationship between plastic waste per capita and mismanaged plastic waste per capita using a scatterplot. Describe the relationship between the two variables.

2. Color the points in the scatterplot by continent. Does there seem to be any clear distinctions between continents with respect to how plastic waste per capita and mismanaged plastic waste per capita are associated?

3. Visualize the relationship between plastic waste per capita and total population as well as plastic waste per capita and coastal population. Do either of these pairs of variables appear to be more strongly linearly associated?

⊕Hint: The colors are from the viridis color palette. Take a look at the functions starting with scale_color_viridis_* in the ggplot2 reference page..

1. Recreate the following plot, and interpret what you see in context of the data. Use the code below to filter out a country with plastic waste per capita over 3 kg/per day. Use the filtered data set for the plot.

plastic_waste <- plastic_waste %>%
  filter(plastic_waste_per_cap < 3)

Submission

Before you wrap up the assignment, make sure all documents are updated on your GitHub repo. we will be checking these to make sure you have been practicing how to commit and push changes.

Once your work is finalized in your GitHub repo, submit the final PDF it to Canvas.

This homework was adapted from Data Science in a Box.