How to make 3D histogram in R

问题

This is my goal: Plot the frequency of y according to x in the z axis.

These are my problems: I have a two columns array (x and y) and need to divide x into classes (p.ex. 0.2 ou 0.5) and calculate the frequency of y for each class of x. The plot should appear like a x-y plot in the "ground" plan and the frequency in the z axis. It could be like a surface or a 3D histogram. I tried to make it using the hist3D function of plot3D package but I don't know what I am doing wrong.

This is an example of what I am trying to do:

https://www.safaribooksonline.com/library/view/r-data-visualization/9781783989508/ch06s05.html

Thanks!!

回答1:

Using some simulated data, this should get you what you want. The key is that you have to create your bivariate bins, accomplished using the cut() function. Then treating the binned factors as levels we can then count the combinations of each factor level using the table() function like below:

library(plot3D)

##  Simulate data:
set.seed(2002)
x <- rnorm(1000)
y <- rnorm(1000)

##  Create cuts:
x_c <- cut(x, 20)
y_c <- cut(y, 20)

##  Calculate joint counts at cut levels:
z <- table(x_c, y_c)

##  Plot as a 3D histogram:
hist3D(z=z, border="black")

##  Plot as a 2D heatmap:
image2D(z=z, border="black")

来源：https://stackoverflow.com/questions/30563340/how-to-make-3d-histogram-in-r