Title: 3D Radial Visualization
Version: 2.3.0
Description: Creating 3D radial visualizations of multivariate data. The package extends traditional radial coordinate visualization (RadViz) techniques to three-dimensional space, enabling enhanced exploration and analysis of high-dimensional datasets through interactive 3D plots. Zhu, Dai & Maitra (2022) <doi:10.1080/10618600.2021.2020129>.
Depends: R (≥ 3.5.0)
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
Encoding: UTF-8
LazyData: true
Imports: rgl (≥ 0.100.19)
Suggests: MixSim, gtools
RoxygenNote: 7.1.1
Author: Yifan Zhu [cre, aut], Fan Dai [aut], Ranjan Maitra [aut], Niraj Kunwar [aut], Gani Agadilov [aut]
Maintainer: Yifan Zhu <yifanzhu@iastate.edu>
NeedsCompilation: no
Packaged: 2025-09-23 07:12:54 UTC; yifanzhu
Repository: CRAN
Date/Publication: 2025-09-30 07:50:02 UTC

G-trasformation function

Description

function to transform discrete or mixture of discrete and continuous datasets to continuous datasets with marginal normal(0,1).

Usage

Gtrans(data, cl = NULL, VariableSelection = FALSE, p_threshold = 0.05, ...)

Arguments

data

The dataset to be transforms. The dataset can be discretein all columns, continuous in all columns or a mixture of continuous columns and discrete columns.

cl

The class information of the dataset. This is not required when VariableSelection = FALSE.

VariableSelection

Logical. If true, anova will be performed to each variable to see whether there is a difference among groups for that variable. The varaible associated with Bonferroni adjusted p-value larger than a threshold will be removed.

p_threshold

The threshold for adjusted p-value in variable selection when VariableSelection = TRUE.

...

Additional arguments passed to internal functions.

Value

A transformed continuous dataset with the same copula as the input dataset and margianl normal(0,1).


Compositions of ancient Chinese celdon pieces

Description

This dataset contains compositional data of ancient Chinese celdon from Longquan and Jingdezhen kiln from North Song to Ming Dynasties.

Usage

celadons

Format

A data frame with 19 variables and 88 observations.

mf

Manufacturer of the celdon piece: FLQ for Jingdezhen and LG for Longquan

era

The manufacturing time and part of the celdon piece in "time-part" format. There are two different parts (body (b) and glaze (g)) and four times (Song Dynasty (S), Yuan Dynasty (Y), Ming Dynastty(M) and Qing Dynasty (QC)).

Al2O3, CaO, CuO, Fe2O3, K2O, MgO, MnO, Na2O, P2O5, PbO2, Rb2O, SiO2, SrO, TiO2, Y2O3, ZnO, ZrO2

The contents of chemical components.


Max-Ratio Projection function

Description

function to project high-dimensional datasets to lower dimention with max-ratio projection.

Usage

mrp(data, cl, nproj = 4, message = TRUE, ...)

Arguments

data

The dataset to apply MRP. Each row is an observation.

cl

The class identification for each observation. The length of cl should be the same as the number of rows of data.

nproj

The number of max-ratio directions to be used in projecting the original data to the projected data.

message

Logical. Wheather to show the accumulative variance explained by the projection directions or not.

...

Additional arguments passed to internal functions.

Value

A list with the elements

projected_df

The projected data with selected number of max-ratio directions.

pccumvar

The cummulative variance explained by the max-ratio principal components.


Overlap matrices for simulated data

Description

This is a list containing three overlap matrices corresponding to the sim_data datasets, showing class separability.

Usage

overlap_mat_sim

Format

A list of 3 matrices, each 5x5, representing overlap between classes


3D Radial Visualization function

Description

3D Radial Visualization function

Usage

radialvis3d(
  data,
  domrp = TRUE,
  doGtrans = FALSE,
  sqrt_scale = FALSE,
  cl = NULL,
  color = NULL,
  pch = 16,
  colorblind = FALSE,
  axes = FALSE,
  point.cex = 1,
  with.coord.labels = TRUE,
  coord.labels = NULL,
  coord.font = 2,
  coord.cex = 1.1,
  with.class.labels = TRUE,
  class.labels = levels(factor(cl)),
  class.labels.locations = NULL,
  opt.anchor.order = FALSE,
  alpha = 0.02,
  lwd = 1,
  axes.col = "black",
  ret.trans = FALSE,
  ...
)

Arguments

data

The dataset to visualize. Each row is an observation.

domrp

Logical. If true, MRP is applied to the origianl dataset. The default number of PCs used is npc = 4.

doGtrans

Logical. If true, Gtrans is applied to the origianl dataset. @seealso Gtrans.

sqrt_scale

Logical. If true, the distance of the points to be visualization will be augmented to squre root of the orginal distance to make points further away from the origin.

cl

The class identification for each observation. The length of cl should be the same as the number of rows of data. If specified, different classes would be visualized with different colors.

color

The colors for different classes. If not specified, rainbow is used.

pch

The point character to be used. It is an integer of a vector of integers of the same length of the nrow of the dataset. See points for a complete list of characters.

colorblind

Logical.The colors for different classes.If true, poits are colorblind friendly.If false, rainbow is used.

axes

Logical.If true, Cartesian axes would be plotted.

point.cex

The size of the data point in RadViz3D. The default value is 1.

with.coord.labels

Logical. If true, labels of coordinates will be added to the visualization.

coord.labels

The labels for components of the dataset. When domrp = TRUE, the coord.labels will be changed to "Xi" representing the the ith direction obtained with MRP.

coord.font

The font for labels of components.

coord.cex

The size of the labels of components.

with.class.labels

Logical. If true, class labels will be added to the visualization.

class.labels

The labels for different classes in the dataset.

class.labels.locations

Locations to put labels for each class. If not specified, an optimal location for each class would be calculated.

opt.anchor.order

Logical. If true, the optimal order of anchor points corresponding to the components would be calculated. This is a very time consuming procedure. Not recommended if the number of components is larger then 6.

alpha

The alpha value that controls the transparency of the sphere in 3d visulization

lwd

The line width in the visualization

axes.col

Colors of the axes, if needed to be displayed

ret.trans

Logical parameter, returns the Radviz3D transformation if TRUE

...

Some other parameters from mrp and Gtrans and rgl functions.

Value

A list with the elements

mrp.res

The result of MRP is the argument domrp = TRUE. See also mrp.

Examples

radialvis3d(data = iris[,-5], cl = iris[,5], domrp = TRUE)

COVID-19 US variants dataset

Description

This is a compositional dataset of the COVID-19 variants in the US from 6/19/2021 to 9/18/2021.

Usage

sarscov2.us.variants

Format

A data frame of 140 observations and 14 variables.

group

The date.

type

weighted

region

Region of the US labelled by numbers.

B.1.1.194, B.1.1.7, B.1.351, B.1.525, B.1.526, B.1.621, B.1.628, B.1.637, Delta, Other*, P.1

COVID-19 variants compositions.


Simulated datasets for testing

Description

This is a list containing three simulated datasets, each with 500 observations and 5 classes, used for testing visualization methods.

Usage

sim_data

Format

A list of 3 data frames, each with 500 observations and 6 variables:

class

Factor with 5 levels representing different classes

X1, X2, X3, X4, X5

Numeric variables with simulated data


Chemical compositions of wine

Description

The dataset contains chemical compositions of wines from 3 cultivars

Usage

wine

Format

A data frame of 178 observations and 14 variables:

cultivar

The cultivar where the wine is produced

Ahl, Ash, Alk, Color, Flvds, Hue, Malic, Mg, Nonfp, ODdil, Phnls, Prol, Pthyns

The content of chemical compositions of the wine