Methods for trend lines
During the previous exercise, you used ggplot2
's default geom_smooth()
.
During this exercise, you will use a glm()
instead.
This will allow you to "see" a logistic regression with ggplot2
.
Specifically, you plot the probability someone takes the bus given their commute distance.
You will need to tell geom_smooth()
to use the glm()
method.
Recall from Chapter 2 that the default family for a glm()
is a Gaussian family, which produces the same results as a lm()
.
Thus, you will also need to specify the method argument from glm()
.
The code for creating gg_jitter
, which you built in the last exercise, has been provided for you.
This exercise is part of the course
Generalized Linear Models in R
Exercise instructions
- Use the
"glm"
method withgeom_smooth()
. - With the
method.args
, set the family to'binomial'
.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Create a jittered plot of MilesOneWay vs Bus2 using the bus dataset
gg_jitter <- ggplot(data = bus, aes(x = MilesOneWay, y = Bus2)) +
geom_jitter(width = 0, height = 0.05) +
ylab("Probability of riding the bus") +
xlab("One-way commute trip (in miles)")
# Add a geom_smooth() that uses a GLM method to your plot
gg_jitter + geom_smooth(method = ___ , method.args = list(___))