How to use make plots

Overview

This vignette describes the usage of the makePlots() function.

makePlots is a function to generate the plots from within the package outside so they can be saved and shared. Plots that have random or non deterministic outcomes have a seed set so they can be reproduced.

makePlots has got the same inputs as pandemonium (see “Data input for pandemonium”) with the exception for settings. This replaces the GUI in the app with one named list to define all settings. Two other small differences in the input to the function are as follows. CovInv is specified for each space. Coordinate functions cannot be passed as a named list instead the functions themselves are passed as getCoordsSpace1 and getCoordsSpace2.

settings

The named list settings has some values that must be specified for every call to makePlots others are optional or only required for some plots. The required values for every call are as follows:

Settings	Description
k=	Numeric, number of clusters
linkage=	Linkage used by stats::hclust
metric=	Metric used by getDists
plotType	Plot to create. Each type and their required settings are described below.

linkage can be one of the following: “complete”, “single”, “average”, “median”, “centroid”, “mcquitty”, “ward.D”, “ward.D2”

metric can be one of the following: “euclidean”, “maximum”, “manhattan”, “canberra”,“binary”, “minkowski”,“user”

Plot specific settings

Each plotType has got its own required settings. for each plot type these are described as follows:

Plot	Description	Required settings
“PC”	Parallel coordinate plot	filt= subset of 1:k (number of clusters) to show in the Plot
“WC”	Plot of clustering on axis from space 2	x=,y= column names of space 2 to be axis showalpha= TRUE or FALSE to show alpha hulls WCa= alpha value for alpha hulls
“chi2”	Plot of score on axis from space 2	x=,y= column names of space 2 to be axis note a function that returns values must be passed in getScores
“sigBins”	Plot of bins on axis from space 2	x=,y= column names of space 2 to be axis note a function that returns bins must be passed in getScores
“heatmap”	Ordered dendrogram plot of clustering	no additional parameters
cluster stats	Plot of cluster stat vs number of clusters	One of the following is what is given as plotType: “within.cluster.ss”, “wb.ratio”,“pearsongamma”, “dunn”, “ch”, “rmax”, “dmax” “dmin”, “dchi2rand”
“Obs”	Plot space 1 variable on axis from space 2	Obs= name of variable from space 1 `x=,y= column names of space 2 to be axis

The following plotTypes are for high dimension views:

“tour”

A detourr tour of the high dimensional space.

Settings	Description
tourspace=	One of “space1” or “space2” to show data of in tour
colouring=	How to colour points in the plot. One of “clustering”, “user”, “bins”, “score”
user_group=	User defined grouping for each point used if “user” is passed for colouring
out_dim=	Dimension of output display. Numeric 2 or 3
tour_path=	Tour path and type to use, one of “grand”,“cmass”,“holes”,“lda”,“pda”,“dcor”,“origin”,“spline”,“radial”,“mahalanobis”,“anomaly”
display=	Display type, one of “scatter”,“slice”
radial_start=	Projection to use as start of radial tour, one of “random”,“cmass”,“holes”,“lda”,“pda”,“dcor”,“origin”,“spline”,“mahalanobis”
radial_var=	Variable to remove by radial tour
slice_width=	Numeric Relative slice volume

“dimRed”

A dimension reduction view of the high dimensional space.

Settings	Description
dimspace=	One of “space1” or “space2” to show data of in tour
colouring=	How to colour points in the plot. One of “clustering”, “user”, “bins”, “score”
user_group=	User defined grouping for each point used if “user” is passed for colouring
algorithm=	Name of algorithm used for dimension reduction
dimReduction	Function for calculating dimension reduction see “using dimension reduction” for more details