prob_var_plot: LOESS Plot for Binary Variable
In wjones127/thesis: Bike Routes Thesis (Title Case)

Description Usage Arguments Details Value Examples

prob_var_plot creates a ggplot2 plot with a Bernoulli LOESS for binary versus continuous variable. This should be useful as a way to determine the relationship between a continuous variable and a binary variable.

1	prob_var_plot(data, x_var, y_var, window_width, method = "bayes")

`data`	The dataframe containing the binary and continuous variable
`x_var`	The name of the continuous variable (as a string)
`y_var`	The name of the binary response variable (as a string)
`window_width`	Width of the window to use for approximating p.
`method`	The method to be used to find confidence intervals for binomial variables. See documentation for binom.confint for options.

It returns a ggplot2 plot, so one could easily add other elements to the plot, such as titles, axes labels, and other themes.

A ggplot2 plot.

x <- rnorm(100, 0, 2)
y <- rbinom(100, 1, invlogit(x))
test_data <- data.frame(x, y)
p <- prob_var_plot(test_data, "x", "y", 2)
p
p + labs(title="Empirical probability versus Gaussian variable")