3D Barplot

3D bar charts are used to provide a 3D look and feel for the data. The third dimension is often used for aesthetic reasons, but it does not improve data reading. Still intended to show comparisons between discrete categories.

Setup

• System Requirements: Cross-platform (Linux/MacOS/Windows)

• Programming language: R

• Dependent packages: plot3D; ggplotify

# Install packages
if (!requireNamespace("plot3D", quietly = TRUE)) {
  install.packages("plot3D")
}
if (!requireNamespace("ggplotify", quietly = TRUE)) {
  install.packages("ggplotify")
}

# Load packages
library(plot3D)
library(ggplotify)

Data Preparation

The loaded data are efficacy data of different doses of drugs in different treatment regimens.

# Load data
data <- read.table("files/Hiplot/003-barplot-3d-data.txt", header = T)

# convert data vector to a matrix
mat <- matrix(rep(1, nrow(data)), nrow = length(unique(data[, 2])))
rownames(mat) <- unique(data[, 2])
colnames(mat) <- unique(data[, 3])
for (i in 1:nrow(mat)) {
  for (j in seq_len(ncol(mat))) {
    mat[i, j] <- data[, 1][data[, 2] == rownames(mat)[i] &
      data[, 3] == colnames(mat)[j]]
  }
}

# View data
mat

       high mid low
Group1   13  25   5
Group2   34   8  25
Group3   21  17  12
Group4   43  10  19

Visualization

# 3D Barplot
p <- as.ggplot(function() {
  hist3D(
    x = 1:nrow(mat), y = seq_len(ncol(mat)), z = mat,
    bty = "g", phi = 20,
    theta = -55,
    xlab = colnames(data)[2],
    ylab = colnames(data)[3], zlab = colnames(data)[1],
    main = "3D Bar Plot", colkey = F,
    border = "black", shade = 0.8, axes = T,
    ticktype = "detailed", space = 0.3, d = 2, cex.axis = 0.3,
    colvar = as.numeric(as.factor(data[, 2])), alpha = 1,
    col = c("#E64B35FF","#4DBBD5FF","#00A087FF","#3C5488FF")
  )

  # Use text3D to label x axis
  text3D(
    x = 1:nrow(mat), y = rep(0.5, nrow(mat)), z = rep(3, nrow(mat)),
    labels = rownames(mat),
    add = TRUE, adj = 0, cex = 0.8
  )
  # Use text3D to label y axis
  text3D(
    x = rep(1, ncol(mat)), y = seq_len(ncol(mat)), z = rep(0, ncol(mat)),
    labels = colnames(mat), bty = "g",
    add = TRUE, adj = 1, cex = 0.8
  )
})

p

Figure 1: 3D Barplot

The 3D histogram shows the different therapeutic effects of different treatment groups (groups 1 to 4) with low, medium and high doses. Group 1 had the best effect with medium dose treatment, group 2 had the best effect with high dose treatment, group 3 had no significant difference with dose treatment, and group 4 had the best effect with high dose treatment.

Tip

Special Parameters:

• Phi: The larger the positive value is, the more it tends to observe from the top down. The greater the negative value, the more likely it is to be observed from the bottom of the three-dimensional figure upwards.
• Theta: The Angle at which a three-dimensional figure rotates horizontally.