Components of a ggplot layer

  1. data - a data frame or tibble
  2. aesthetic mappings (aes) - specify the variables to be plotted
  3. geom - specify the geometric to be drawn in the layer (e.g. points, lines, bars, etc.)
  4. stat - compute and plot statisitcal transforms of the raw date
  5. position - adjustments for overlapping elements
  6. facet - generate multiple related plots

ggplots are layered compositions of these various components. To add components to the plot we simply add them with a “+”

Template for ggplot components

p1 <- ggplot(data= <DATA>,mapping=aes(<MAPPINGS>) + 
<GEOM_FUNCTION>(aes(<MAPPINGS>),
                stat=<STAT>,
                position=<POSITION>) +
                <COORDINATE_FUNCTION> +
                <FACET_FUNCTION>

print(p1)
ggsave(plot=p1, filename="MyPlot",width=5,height=3,units="in",device="svg")

Packages

library(tidyverse)
## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.2 ──
## ✔ ggplot2 3.3.6      ✔ purrr   0.3.4 
## ✔ tibble  3.1.8      ✔ dplyr   1.0.10
## ✔ tidyr   1.2.1      ✔ stringr 1.4.1 
## ✔ readr   2.1.2      ✔ forcats 0.5.2 
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
library(ggplot2) 
library(ggthemes) #extended themes for ggplot
library(patchwork) #tools for composing multiple panels

For these examples we will use the mpg built in data set from ggplot2

?mpg

Quick plotting for data exploration with qplot

# simple histogram
qplot(x=mpg$cty) #y-axis is count of observation with a given value of the x
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

We can modify the colors:

qplot(x=mpg$cty,fill=I("goldenrod"),color=I("black")) 
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Here we used I() because if we directly call the color name, R will look for a variable with that name and attempt to apply colors based on the values of the variable.

# simple density plot
qplot(x=mpg$hwy,geom="density")

# simple scatter plot with linear regression line
qplot(x=mpg$cty,y=mpg$hwy,geom=c("point","smooth"), method="lm")
## `geom_smooth()` using formula 'y ~ x'

We can color points by a variable

qplot(x=mpg$cty,y=mpg$hwy, col = mpg$class, geom=c("point"))

# The regression is now applied to each class separately
qplot(x=mpg$cty,y=mpg$hwy, col = mpg$class, geom=c("point","smooth"), method="lm")
## `geom_smooth()` using formula 'y ~ x'

# basic boxplot
qplot(x=mpg$fl, y=mpg$cty,geom="boxplot",fill=I("green"))

Interpretation:

Interquartile range (IQR)

# simple barplot 
qplot(x=mpg$fl,geom="bar")

Most often bar charts are used to display summary values by group. One way to do this is to use dplyr to calculate summary stats and plot those with col

# bar plot with specified means

mpg_summary<-mpg %>%
  group_by(class) %>%
  summarise(mean_hwy = mean(hwy))

p<-qplot(x=mpg_summary$class, y=mpg_summary$mean_hwy, geom="col")

p

We can flip this plot by layering coord_flip:

p + coord_flip()

Full ggplots

Themes and fonts

p1 <- ggplot(data=mpg, mapping=aes(x=hwy,y=cty)) +
  geom_point()
print(p1)

Now that we have created the plot object p1 we can layer elements and themes onto it

 p1 + theme_bw() #clean with gridlines

p1 + theme_classic() # no grid lines

p1 + theme_dark() # good for brightly colored points

 p1 + theme_void() # shows data only

If you must:

 p1 + theme_excel()

Theme modifications

Control font and size of text

 # use theme parameters to modify font and font size. 
 p1 + theme_classic(base_size=25,base_family="sans")

base_<XXX> modifies throughout the figure. You can also control specific elements:

p1 + labs(x="Highway Gas Milage (mpg)") + # replace the default label on the y-axis
  theme_classic() +
  theme(axis.title.x = element_text(size = 15, angle = 45, vjust = 0.5)) # specifically change font size, angle, and position of x-axis label

Minor modifications

p1 <- ggplot(data=mpg) +
  aes(x=hwy,y=cty) + 
  geom_point(size=7,
             shape=25,
             color="black",
             fill="hotpink") +
  labs(title="My graph title here",
       subtitle="An extended subtitle that will print below the main title",
       x="My x axis label",
       y="My y axis label") +
  xlim(0,50) +
  ylim(0,50)
      
p1

Shape options for points

A convenient document with base colors can be found here

colors() # will return a list of all colors in base R
Lesson developed with materials adapted from: https://gotellilab.github.io/Bio381/