A fuzzy monster in a beret and scarf, critiquing their own column graph on a canvas in front of them while other assistant monsters (also in berets) carry over boxes full of elements that can be used to customize a graph (like themes and geometric shapes). In the background is a wall with framed data visualizations. Stylized text reads “ggplot2: build a data masterpiece.

Figure from Allison Horst

Introduction

We will will build upon our last lesson on ggplot101 which focused on an overall understanding of the grammar of graphics, basic syntax, adding data, aesthetic mappings, and geoms. Today we will focus on some of the other more commonly adjusted layers:

Load libraries and data

Before we get started, let’s load our libraries and data.

library(tidyverse)
library(gardenR)

And let’s remember whats in garden_harvest.

glimpse(garden_harvest)
## Rows: 781
## Columns: 5
## $ vegetable <chr> "lettuce", "radish", "lettuce", "lettuce", "radish", "lettuc…
## $ variety   <chr> "reseed", "Garden Party Mix", "reseed", "reseed", "Garden Pa…
## $ date      <date> 2020-06-06, 2020-06-06, 2020-06-08, 2020-06-09, 2020-06-11,…
## $ weight    <dbl> 20, 36, 15, 10, 67, 12, 9, 8, 53, 19, 14, 10, 48, 58, 8, 121…
## $ units     <chr> "grams", "grams", "grams", "grams", "grams", "grams", "grams…

Facets

Faceting allows to create small multiples of plots, enabling the easy comparison across the entirety of your data. A benefit of plots like this is they are all structured the same way, so once you understand one, you can begin to look at trends across groups/treatments/conditions simply and easily.

Here is a more infographic example of using small multiples.

A plot of small multiples, laid out in the shape of the United States, to show how life expectancy differs from the average in the US across the states. There is a relatively low life expectancy in Washington DC, and the south, and relatively high in Hawaii, the northeast, upper midwest, and west.

Figure from Five Thirty Eight

So we can easily see that states with more of a maroon color have a lower than average life expectancy, while those that are higher than average are orange. We also can see easily where each state is on the map, so we can begin to understand how geography is related to life expectancy. We can also see which states have gotten better (i.e. their people live longer) with time, and those that haven’t. And this is all with a quick glance!

If we look back to the plot we were using as our example last week, can see how we have a plot faceted by tomato variety.

First lets select only the data for tomatoes.

# filter data to include only tomatoes 
# filter() is a useful function from dplyr (part of tidyverse)
# it allows us to select observations based on their values
garden_harvest_tomato <- garden_harvest %>%
  filter(vegetable == "tomatoes")

Let’s remember what our base plot is currently looking like.

garden_harvest_tomato %>%
  ggplot(aes(x = date, y = weight, color = variety)) +
  geom_line() +
  geom_point(size = 1)

See how crowded this is? I think faceting might help us better see our data by variety.

There are two functions that allow you to facet:

  • facet_wrap: allows to lay out your facets in a wrapped type. You can use facet_wrap if you have 1 variable you’d like to facet on.
  • facet_grid: allows you to lay out your facets in a grid. You can use facet_grid if you have 1 or 2 variables you’d like to facet on.

There are a few different sets of syntax that work for faceting, but I think this is the most intuitive.

garden_harvest_tomato %>%
  ggplot(aes(x = date, y = weight)) +
  geom_line() +
  geom_point(size = 1) +
  facet_wrap(vars(variety))

We will get a very reasonably different looking plot with facet_grid with the default settings.

garden_harvest_tomato %>%
  ggplot(aes(x = date, y = weight)) +
  geom_line() +
  geom_point(size = 1) +
  facet_grid(vars(variety))

Note because you have provided only one variable, ggplot has put that facet in one row.

garden_harvest_tomato %>%
  ggplot(aes(x = date, y = weight)) +
  geom_line() +
  geom_point(size = 1) +
  facet_grid(cols = vars(variety))

We can make the faceting go by column, but this is also looks bad in this case

However, you might be thinking now that if you have two variables, and you want to facet by the combination of them, you could do that with facet_grid. Here is an example with the mpg dataset from the tidyverse (since there isn’t really good data to demonstrate this from garden_harevst).

mpg %>%
  ggplot(aes(x = cty, y = hwy)) + # city and highway gas mileage
  geom_point() +
  facet_grid(cols = vars(class), # category of car
             rows = vars(drv)) # type of drive train, 4 wheel, front, rear

The default in both facet_wrap and facet_grid are for the x and y-axis to be fixed and constant among all the plots. This is often what you want to take advance of the comparisons between small multiples, but this is something you can change if you want. You can adjust the scales within facets to:

  • scales = "fixed": both the x- and y-axes are fixed for all plots to be the same (this is the default)
  • scales = "free": both the x- and y-axes are set per plot
  • scales = "free_x": the x-axis scales are free between plots
  • scales = "free_y": the y-axis scales are free between plots
garden_harvest_tomato %>%
  ggplot(aes(x = date, y = weight)) +
  geom_line() +
  geom_point(size = 1) +
  facet_wrap(vars(variety), scales = "free")