Back to schedule

acitelli <- read.csv("acitelli.csv")

There are quite a few ways to make figures in R, we’ll the popular package ggplot2. You can find a cheat sheet for ggplot2 here. Be sure to install it first if you never have, or if you need to update it.

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

Visualizing Data with qplot()

Quick Histogram

First, let’s make a histogram for satisfaction. The easiest way to make a figure with ggplot2 is with the qplot() function. This stands for quick plot. Notice in the code below that we did not specify anything about a histogram. qplot() guesses which type of plot we want based on the variable’s type (i.e., integer, number, double, factor, character).

qplot(satisfaction, data = acitelli)

There are too many bins (it defaults to 30 bins), we can ask for a specific number by adding the bins = argument. Try playing around the bin number below to find the optimal plot. I put 10 in there as a placeholder.

#play around with different numbers of bins
qplot(satisfaction, data = acitelli, bins = 10)

Remember that you can run the ? code to see the help file for any function.

?qplot

We might want to see if the distributions are different for men and women. We can do this by mapping Gender to the fill aesthetic. Note that we could used color = if we want a hollow histogram. Try it out!

qplot(x = satisfaction, fill = as.factor(gender), data = acitelli, bins = 10)

Why did we have to change the variable type of gender? What happens if you don’t?

Quick Density Plot

An alternative to the histogram is the density plot. It displays a smoothed distribution and the area under the curve always sums to 1, thus, it’s good for comparing two groups with different n’s.

qplot(x = satisfaction, fill = as.factor(gender), data = acitelli, alpha = .05, geom = "density")

What would you say the alpha = argument is doing? What happens if you take it away? What if you change the value?

#play with alpha here

Quick Scatterplot

We might also want a scatter plot. Again,qplot() guesses what we want, but it’s a good idea to specify which variable goes on the x-axis and which goes on the y-axis.

qplot(x = tension, y = satisfaction, data = acitelli)

We can even add a third variable, mapping it to color.

qplot(x = tension, y = satisfaction, color = gender, data = acitelli)

What’s going on with the color scale for gender? Can you explain why this happend? What happens if you change the variable type of gender?

Quick Boxplots

We can ask for side-by-side boxplots when our x variable is categorical. In this case qplot() does NOT know what to do, so we tell it we want boxplots by adding the argument geom = "boxplot".

qplot(y = satisfaction, x = as.factor(gender), data = acitelli, geom = "boxplot")

What would happen if you added fill = as.factor(gender) as an argument. Try it by adding a new chunk below.

ADD A NEW CHUNK HERE

Full Blown ggplots with ggplot()

Histogram

For more complex figures we will need to move away from using the qplot() function in favor of the heavy duty ggplot() function. To get a sense of how ggplot() builds plots, first we will just run the empty function.

ggplot()

Next, we can add the data and start mapping variables to aesthetics.

ggplot(acitelli, aes(x = satisfaction))

After we have specified aesthetic mappings, we can then add geoms. Notice that we make use of the + symbol with ggplots. The + needs to be on the right of each piece of the plot. We add a histogram with the geom_histogram() function.

ggplot(acitelli, aes(x = satisfaction)) +
  geom_histogram()

Note in the histogram above the y-axis are the counts of observations in each bin. There were some calculations involved in getting these counts. Counts are the default statistic when you ask for a histogram. We can change the number of bins by adding bins = inside of the geom_histogram() function.

ggplot(acitelli, aes(x = satisfaction)) +
  geom_histogram(bins = 12)

Just as before, if we want overlayed histograms, we can map gender to the fill aesthetic.

ggplot(acitelli, aes(x = satisfaction)) +
  geom_histogram(aes(fill = as.factor(gender)), bins = 12)

Alternatively, we can ask for separate facets for each level of the Gender variable with the fact_wrap() function. Notice that there is a ~ before Gender inside of this function.

ggplot(acitelli, aes(x = satisfaction)) +
  geom_histogram(bins = 12) +
  facet_wrap(~as.factor(gender))

Scatter Plot

Next we’ll make that scatter plot again. We’ll map tension to the x-axis and satisfaction to the y-axis. Then we’ll add geom_point().

ggplot(acitelli, aes(x = tension, y = satisfaction)) +
  geom_point()

Why does it appear as though there is far less data than there really is? Check out the plot when we use geom_jitter(). What do you think geom_jitter() does?

ggplot(acitelli, aes(x = tension, y = satisfaction)) +
  geom_jitter(height = 0, width = .05, alpha = 0.5)

We can add more than one geom. To the jittered scatter plot we can add a least squares regression line with geom_smooth(). Inside of geom smooth we need to specific method = "lm", the lm stand for linear model. We can also turn off the standard errors with se = 0.

ggplot(acitelli, aes(x = tension, y = satisfaction)) +
  geom_jitter(height = 0, width = .05, alpha = 0.5) +
  geom_smooth(method = "lm", se = 0)

Again, we can map gender to the color aesthetic. Note: Here the mapping of gender to color is done in the geom_smooth() function—but what would happen if you moved it to the aes() function in the ggplot() function?

ggplot(acitelli, aes(x = tension, y = satisfaction)) +
  geom_jitter() +
  geom_smooth(aes(color = as.factor(gender)), method = "lm", se = 0)

Or use facet_wrap(). In the code below, what does color = "salmon" do? Could you change the regression lines to salmon?

ggplot(acitelli, aes(x = tension, y = satisfaction)) +
  geom_jitter(color = "salmon") +
  geom_smooth(method = "lm", se = 0) +
  facet_wrap(~as.factor(gender))

Labels and Colors

We can create a plot object with the <- symbol. Then to print the plot we’d need to run a line with the name of our plot.

myplot <- ggplot(acitelli, aes(x = tension, 
                               y = satisfaction, 
                               color = as.factor(gender))) +
  geom_jitter(height = 0, width = .05) +
  geom_smooth(method = "lm", se = 0)

myplot

Then, we can add to that plot object. We can add x labels, x labels, change the colors, and the theme. There is much more that you can do with ggplot2!

myplot + 
  xlab("Tension") +
  ylab("Satisfaction") +
  scale_color_manual(values = c("gold", "dodgerblue"), 
                     name = "Gender") +
  ylim(1,4) +
  xlim(1,4) +
  theme_minimal()

Back to schedule