• 3. A continuous variable

Motivating scenario: You want to see the distribution of a single continuous variable.

Learning goals: By the end of this sub-chapter you should be able to

  1. Familiarize yourself with making plots using the ggplot2 package.
  2. Use geom_histogram() to make a histogram.
    • Specify the width (with binwidth) or number (with bins) of bins in a histogram.
  3. Use geom_density() to make a density plot.
  4. Change the outline color (with color) and the fill color (with fill).

Visualizing Distributions

Let’s first consider the distribution of a single continuous variable—pollinator visitation. There are some natural questions we would want answered early on in our analysis:

  • Are most flowers visited frequently, or do visits tend to be rare?
  • Is the distribution symmetric, or is it skewed, with many flowers receiving few or no visits and a small number receiving many?

As we will see throughout the term, understanding the shape of our data helps guide our analysis. Let’s start with two common visualizations for distributions:

  • A histogram bins the x-axis variable (in this case, visitation) into intervals and shows the number of observations in each bin on the y-axis. This allows us to see how frequently different levels of visitation occur.
  • A density plot fits a smooth function to the histogram, providing a continuous representation of the distribution. This smoothing can sometimes make patterns easier (or harder) to see. Later, we’ll see that density plots can also help compare distributions.

Making histograms & density plots: A Step-by-Step Guide. To create these visualizations in ggplot2, we need to (minimally) specify three key layers:

  1. The data layer: This is the dataset we’re plotting—in this case, ril_data. We pass this as an argument inside the ggplot() function, e.g., ggplot(data = ril_data).
  2. The aesthetic mapping: This tells R how to map each variable onto the plot. In a histogram, we map the variable whose distribution we want to visualize (in this case, mean_visits) onto the x-axis. We define this inside the aes() argument within ggplot(), e.g., ggplot(data = ril_data, aes(x = mean_visits)).
  3. The geometric object (geom) that displays the data: In this case, we use geom_histogram() for a histogram or geom_density() for a density plot. These are added to the plot using a +, as shown in the examples below.

Inside the ggplot() function, we tell R that we are working with the ril_data, and we map mean_visits onto x inside aes(). The result is a boring canvas, but it is the canvas we build on.

ggplot(data = ril_data, aes(x = mean_visits))

But now we want to show the data. To do so we need to add our geom. In this case we show the data as a histogram.

The resulting plot shows that most plants get zero visits, but some get up to five visits.

ggplot(ril_data,aes(x = mean_visits))+
  geom_histogram()

To take a bit more control of my histogram, I like to

  • Specify the number or width of bins (with the binwidth and bins arguments, respectively).
  • Add white lines between bins to make them easier to see (with color = "white").
  • Spruce up the bins by specifying the color that fills them (with e.g. fill = "pink").
ggplot(ril_data, aes(x = mean_visits))+
  geom_histogram(binwidth = .2, color = "white", fill = "pink")

Now you can easily make a density plot if you prefer!

ggplot(ril_data, aes(x = mean_visits))+
  geom_density(color = "white", fill = "pink")

Interpretation: Returning to our motivating questions, we see that most plants receive no visits, but this distribution is skewed with some plants getting lots of visits.