nice3DPlot: Plot a Nice 3D Plot

In ryannick28/CustomFunctionsYrotha: Collection of my custom functions

Plot a Nice 3D Plot

Description

This function creates a nice plot of 3D-data, including a visualization of the density. If no data is supplied, then only the coordinate system is drawn, which can be used for further plotting. If data is provided, it must be numeric. One can select if only the points, only the density or both shall be plotted. One can also use it to plot simulated data (normally distributed). In that case one has to provide a covariance matrix and a vector of the means.

Usage

nice3DPlot(X = NULL, whatToPlot = c("P","D", "PD"),
plotFit = c('no','lin','int','int2','int3'), catVar = factor(1),
covMat = NULL, means = NULL, nSim = 500, pointCol = 1, colRamp = NULL,
pointSize = NULL, spheres = FALSE, pointTrans = 0.8, Dtransp_fac = 0.06,
colres = 50, gridRes = 30, h = 0.3, DpointSize = 30, axesNames = NULL,
axesTicks = FALSE, gridCol = "grey", axesLeng = NULL, zoom = 1,
add = FALSE, htmlFilename = NULL, ...)

Arguments

X	The data to be plotted. This should either be a matrix or a dataframe. The function will use the first three columns for plotting, assigning them the axes x, y, z (in this order).
whatToPlot	A string defining what to plot. One can either plot only the data points ('P'), only the density ('D') or both ('PD').
plotFit	Add a surface showing the fit of a linear regression (use 'lin' for model without interaction and 'int' for model with interaction). Default is 'no' for no drawn fit. The options 'int2' and 'int3' are only for cases when a factor is provided with catVar. If such a factor is provided it is included in the displayed model fit. The 'lin' option displays a model in which the factor is included as a fixed factor. The 'int' option displays a model with an interaction between the numeric predictors and the factor of catVar included as a fixed factor. The 'int2' option displays a model in which the interaction terms between the numeric predictors and the factor of catVar are included. The 'int3' option displays a model in which all interaction terms are included.
catVar	An optional factor. Providing such a factor will automatically colour the points according to factor levels. Additionally, the factor will be included when plotting the fit of a linear model with the plotFit argument. See the plotFit argument for details.
covMat	Optional covariance matrix (has to have 3rows/columns, order: x,y,z) to plot simulated data from a multivariate normal distribution.
means	A vector of length three with the x, y and z mean values of the simulated data.
nSim	Number of simulated data points.
pointCol	Colour of points.
colRamp	A numeric vector of the same length as there are observations in X. It contains the values which will determine the colour of points along a blue-red colour continuum.
pointSize	Size of points.
spheres	Logical indicating whether rgl.spheres() should be used for drawing points (instead of rgl.points()). Depending on hardware (apparently) rgl.points() may display datapoints as boxes.
pointTrans	Transparency of points.
Dtransp_fac	Determines the transparency of the density pixels. With smaller values the displayed density will be more transparent.
colres	Determines the number of steps in the colour progression of the density pixels.
gridRes	Resolution of the grid for which the density values are calculated.
h	Bandwidth of the 3D-Kernel Density Estimator. Changing the value of h can often lead to drastically improved visualization of the density.
DpointSize	Size of the points of the density grid.
axesNames	An optional vector of length 3, giving the axes names of the x, y and z axis (in this order).
axesTicks	Optional logical value to add ticks to the axes.
gridCol	Color of the floor grid in the 3D-coordinate system.
axesLeng	Manually set axes length. Will be extended to a vector of length three assigning the lengths of the x, y and z axis (in this order).
zoom	How much to zoom into the graph (so far only possible with origin as center). Note: On my machine the field of view can be adapted by moving cursor while having the right-click pressed.
add	Set to TRUE in order to add points to an existing 3D Plot.
htmlFilename	A string giving the name/path of a html file to which the 3D plot will be saved to. If NULL than no html file is created. Dont forget to add the ".html" to the name.
...	Additional arguments which will be passed to the rglwidget() function, which is responsible for creating a html version of the 3D plot.

Details

The 3D-density is created by generating a grid of points around the data and colouring each point of the grid according to the density value at that position. In addition, the transparency of the gridpoints is set according to the density values.

Examples

# Only coordinate-system:
nice3DPlot()

# Plot data with density:
nice3DPlot(X = data.frame(x=rnorm(100), y=rnorm(100), z=rnorm(100)),
           whatToPlot = 'PD', h = 0.7, DpointSize = 30, zoom=0.9)

# Simulated data:
m <- rep(0,3)
covM <- matrix(c(1,0.5,0.5,0.5, 1,0.5,0.5,0.5,5), ncol=3)
nice3DPlot(covMat = covM, means = m, whatToPlot = 'PD', h = 0.7,
           DpointSize = 30, zoom=1.5, axesLeng = 5, pointTrans = 0.5)

ryannick28/CustomFunctionsYrotha documentation built on June 1, 2025, 4:02 p.m.