The plot3d package can be used to generate stunning 3-D plots in R. It can generate an interesting array of plots, but in this recipe we will focus on creating 3-D scatterplots. These arise in situations where we have three variables, and we want to plot the triplets of values on the x-y-z space.

We will generate a dataset containing random Gaussian numbers for three variables, and we will plot them into the same plot using the plot3d package.

This package can be installed in the usual way via install.packages("plot3D").

1. Import the plot3D library:

library(plot3D)

2. Generate a dataset containing random Gaussian numbers for three variables:

x = rnorm(100)
y = rnorm(100)
z = x + y + rnorm(100,0,0.3)
idrow = 1:100

3. Plot the variable in the same plot:

scatter3D(x, y, z, bty = "g", colkey = TRUE, main ="x-y-z plot",phi = 10,theta=50)
text3D(x, y, z, labels = idrow, add = TRUE, colkey = FALSE, cex = 0.5)

The following screenshot is the resulting 3D plot:

The scatter3D function draws the scatterplot, and we have an interesting set of options for it. We can turn the color key on/off using the colkey parameter. phi and theta control the angles that will be used to show the plot. The color key is quite useful as it helps to highlight the observations that have higher Z values. This is useful because in 3-D plots it is sometimes difficult to understand a single image without rotating it. We are also using the text3D function to print the values for Z for each point. This step could certainly be omitted, but it is generally useful for isolating individual observations.