Install & load ggplot2()

install.packages("ggplot2")
library(ggplot2)

Practice data

The mtcars dataset is part of the base R package, and is the daataset that we will use to explore ggplot2.

Sample plot

Use the following code to make a scatterplot of horsepower(hp) x miles per gallon(mpg).

ggplot(data=mtcars,mapping = aes(x=hp, y=mpg))+
  geom_point()

Basic commands & necessary arguments.

• We initiate a plot with the ggplot() command.
• Then, we include a geom command to define the geometric objects that we will use to visualize the data.
• These two commands are joined together by a “+”

There are two major arguments which must be defined within each ggplot & geom command, otherwise they will not produce any output.

• data=

  • This is where you will define the data that is to be referenced by the plot(s).

• mappings=

  • This is where you will define the visual properties, or aesthetic mappings, of the plot(s).

Placement of the data= and mapping= arguments

1. You can place both arguments within the top-level ggplot() command. This will tell ggplot to apply this data & aesthetic mapping to all geoms that follow.

Notice that both the point geom, and the line geom are illustrating the same thing (hp x mpg).

ggplot(data=mtcars,mapping = aes(x=hp, y=mpg))+
  geom_point()+
  geom_line()

2. You can place the arguments within each geom() command. This will tell ggplot to apply this data & aesthetic mapping only to this geom.

Notice below that the geom_line is not plotted, because there is no data & no aesthetic mapping defined

ggplot()+
  geom_point(data=mtcars,mapping = aes(x=hp, y=mpg))+
  geom_line()

3. You can pass the data= argument at the ggplot() level, and the aesthetic mapping argument at the geom level. This results in all geoms using the same data, but each geom having uniquely defined aesthetic mappings.

Notice below that geom_point() shows horsepower by miles per gallon, while geom_line() shows horsepower by quarter mile time.

ggplot(data=mtcars)+
  geom_point(mapping = aes(x=hp, y=mpg))+
  geom_line(mapping = aes(x=hp, y=qsec))

Data-defined vs. Manually-defined Aesthetics

Any mapping arguments inside of the aes() command are tied to the data. Mapping arguments outside of the aes() command pertain to the entire geom, or ggplot object.

Example(Left): Cars with Manual transmission are colored blue. Size indicates number of cylinders.

ggplot(data=mtcars)+
  geom_point(aes(x=hp, y=mpg, color=am,size=cyl))

Example(Right): Size indicates number of cylinders, however all points are blue since color is not tied to values in the data. Again, this is achieved by putting the color argument outside of the aes() command.

ggplot(data=mtcars)+
  geom_point(aes(x=hp, y=mpg,size=cyl),color="blue")

## 
## Attaching package: 'gridExtra'

## The following object is masked from 'package:dplyr':
## 
##     combine

Contiuous vs. Categorical values

Notice that we have a color scale legend for the am variable. That is because it is being treated like a continuous variable. We can convert it to a factor with the as.factor() command in order to have it treated categorically.

Example(Left): “am” treated as a continuous variable

ggplot(data=mtcars)+
  geom_point(aes(x=hp, y=mpg, color=am,size=cyl))

Example(Right): “am” converted to factor and treated as a categorical variable.

ggplot(data=mtcars)+
  geom_point(aes(x=hp, y=mpg, color=as.factor(am),size=cyl))

Note: repeating the “as.factor()” command can make for bulky code very quickly. If we will be consistently treating “cyl” and “am” as factors, we can define a new dataframe and re-classify these variables the way we want them.

library(dplyr)
newmtcars <- mtcars%>%
  mutate(cyl = as.factor(cyl),
         am = as.factor(am))

ggplot(data=newmtcars)+
  geom_point(aes(x=hp, y=mpg, color=am,size=cyl, shape=am))

Color & Fill

• color refers to the outline of a geometric object
• fill* refers to the color that covers the surface area of an object.

Example(Left): “color”

ggplot(data=newmtcars)+
  geom_bar(aes(x=am, color=am))

Example(Right): “fill”

ggplot(data=newmtcars)+
  geom_bar(aes(x=am, fill=am))

aes(x=x,y=y, color=color) vs. aes(x=x,y=y),color=color

Example(Left): When color/fill is defined inside of the aes() command, the color/fill assignments will correspond to the data.

ggplot(data=newmtcars)+
  geom_bar(aes(x=am, fill=am))

Example(Right): When color/fill is defined outside fo the aes() command, the color/fill assignment applies to the entire plot & is manually-defined.

ggplot(data=newmtcars)+
  geom_bar(aes(x=am), fill="aquamarine1")

Here is a great reference for colors that are recognized by name in r.

Position

There are arguments that can be added to ggplot() to adjust the position of your geoms.

• position=“jitter”

• position=“fill”

• position=“dodge”

• position=“stack”

• position=“identity”

ggplot()+
  geom_point(data=mtcars, aes(x=cyl, y=mpg))

ggplot()+
  geom_point(data=mtcars, aes(x=cyl, y=mpg),position="jitter")

ggplot(data=mtcars)+
  geom_point(aes(x=cyl, y=mpg),position=position_jitter(w = 0.1, h = 0.1))

ggplot(mtcars)+ 
  geom_bar(aes(x= factor(cyl), fill = factor(am)))

ggplot(mtcars)+ 
  geom_bar(aes(x= factor(cyl), fill = factor(am)), position = "fill")

ggplot(mtcars, aes(factor(cyl), fill = as.factor(am))) +
  geom_bar()

ggplot(mtcars, aes(factor(cyl), fill = as.factor(am))) +
  geom_bar(position = "dodge")

Size

You can manually or dynamically assign size values to your points.

Example(Left):Size manually assigned (outside of the aes() command).

ggplot()+
  geom_point(data=mtcars, aes(x=cyl, y=mpg, color=cyl),size=5, position="jitter")

Example(Right):Size dynamically assigned (inside of the aes() command).

ggplot()+
  geom_point(data=mtcars, aes(x=cyl, y=mpg, color=cyl, size=cyl),position="jitter")

Shape

Your points can take on various shapes which can, again, be assigned manually or dynamically.

Example(Left):Shape manually assigned (outside of the aes() command).

ggplot(mtcars)+
  geom_point(aes(x=cyl, y=mpg),shape=3, position="jitter")

Example(Right):Shape dynamically assigned (inside of the aes() command).

ggplot(data=mtcars)+
  geom_point(aes(x=cyl, y=mpg, shape=factor(cyl)), position="jitter")

The numeric assignments of each primary shape for manual selection are as follows.

geom_point

ggplot(mtcars)+
  geom_point(aes(x=hp, y=mpg))

geom_bar & geom_col

Example(Left):geom_bar uses stat=“count” as its default, plotting the frequencies of each x-axis value.

ggplot(starwars)+
  geom_bar(aes(x=gender))

Example(Right):geom_col uses stat=“identity” as its default, plotting the summed y-values for each x-value.

ggplot(starwars)+
  geom_col(aes(x=gender, y=height, fill=gender))

geom_histogram

ggplot(mtcars)+
  geom_histogram(aes(x=hp))

geom_line & geom_smooth

Example (Left): geom_line produces a simple line graph which follows x and y values exactly.

ggplot(mtcars)+
  geom_line(aes(x=hp, y=mpg))

Example (Right): geom_smooth produces a line graph which is smoothed out according to a method= argument.

ggplot(mtcars)+
  geom_smooth(aes(x=hp, y=mpg))

## `geom_smooth()` using method = 'loess'

geom_tile

Example: geom_tile is useful for visualizing matrixes and effectively producing “heatmaps”. In the example below, geom_tile simplifies the interpretation of temperature readings across a 5 month period.

ggplot(airquality, aes(x = Month, y = Day)) +
  geom_tile(aes(fill=Temp))+
  scale_fill_gradient(name = 'Temperature', low = 'white', high = 'red')+
  theme(legend.position="none")+
  xlab("")+ylab("")+labs(title="Temperature Readings by Date") 

geom_boxplot

ggplot(mtcars)+
  geom_boxplot(aes(x=am, y=mpg, group=am))

geom_text