Introduction to ‘parallelPlot’

library(parallelPlot)

Basic usage (dataset uses factor type)

parallelPlot(iris)

‘species’ column is of factor type and has box representation for its categories.

refColumnDim argument (referenced column is categorical)

parallelPlot(iris, refColumnDim = "Species")

Each trace has a color depending of its ‘species’ value.

categoricalCS argument

parallelPlot(iris, refColumnDim = "Species", categoricalCS = "Set1")

Colors used for categories are not the same as previously (supported values: Category10, Accent, Dark2, Paired, Set1).

refColumnDim argument (referenced column is continuous)

parallelPlot(iris, refColumnDim = "Sepal.Length")

Each trace has a color depending of its ‘Sepal.Length’ value.

continuousCS argument

parallelPlot(iris, refColumnDim = "Sepal.Length", continuousCS = "YlOrRd")

Colors used for lines are not the same as previously. Supported values: “Viridis”, “Inferno”, “Magma”, “Plasma”, “Warm”, “Cool”, “Rainbow”, “CubehelixDefault”, “Blues”,“Greens”, “Greys”, “Oranges”, “Purples”, “Reds”, “BuGn”, “BuPu”, “GnBu”, “OrRd”, “PuBuGn”,“PuBu”, “PuRd”, “RdBu”, “RdPu”, “YlGnBu”, “YlGn”, “YlOrBr”, “YlOrRd”)

Basic usage (dataset doesn’t use factor type)

parallelPlot(mtcars)

Several columns are of numerical type but should be of factor type (for example ‘cyl’).

categorical argument

categorical <- list(NULL, c(4, 6, 8), NULL, NULL, NULL, NULL, NULL, c(0, 1), c(0, 1), 3:5, 1:8)
parallelPlot(mtcars, categorical = categorical, refColumnDim = "cyl")

‘cyl’ and four last columns have a box representation for its categories.

categoriesRep argument

categorical <- list(NULL, c(4, 6, 8), NULL, NULL, NULL, NULL, NULL, c(0, 1), c(0, 1), 3:5, 1:8)
parallelPlot(mtcars, categorical = categorical, refColumnDim = "cyl", categoriesRep = "EquallySizedBoxes")

Within a category column, the height assigned to each category can either be equal for each category (EquallySizedBoxes) or calculated to reflect the proportion of lines passing through each category (EquallySpacedLines).

arrangeMethod argument

categorical <- list(NULL, c(4, 6, 8), NULL, NULL, NULL, NULL, NULL, c(0, 1), c(0, 1), 3:5, 1:8)
parallelPlot(mtcars, categorical = categorical, refColumnDim = "cyl", arrangeMethod = "fromLeft")

Within a category box, the position of lines is computed to minimize crossings on the left of the box. arrangeMethod can also be set to fromRight to minimize crossings on the left of the box (default behavior). ‘fromBoth’ allows to merge the two behaviors (see next example). To turn this ordering off (for example for performance reasons), arrangeMethod can also be set to fromNone.

arrangeMethod argument (using fromBoth)

categorical <- list(NULL, c(4, 6, 8), NULL, NULL, NULL, NULL, NULL, c(0, 1), c(0, 1), 3:5, 1:8)
parallelPlot(mtcars, categorical = categorical, refColumnDim = "cyl", arrangeMethod = "fromBoth")

Within a category box, lines have an incoming point and an outgoing point; these points are ordered to minimize crossings on the left and on the right of the box.

inputColumns argument

categorical <- list(NULL, c(4, 6, 8), NULL, NULL, NULL, NULL, NULL, c(0, 1), c(0, 1), 3:5, 1:8)
inputColumns <- c(FALSE, TRUE, TRUE, FALSE, FALSE, TRUE, FALSE, TRUE, TRUE, TRUE, TRUE)
parallelPlot(mtcars, categorical = categorical, inputColumns = inputColumns, refColumnDim = "cyl")

The column name is blue for outputs and green for inputs (in shiny mode, inputs can be edited).

histoVisibility argument

histoVisibility <- rep(TRUE, ncol(iris))
parallelPlot(iris, histoVisibility = histoVisibility)

An histogram is displayed for each column.

invertedAxes argument

invertedAxes <- rep(FALSE, ncol(iris))
invertedAxes[2] <- TRUE
parallelPlot(iris, invertedAxes = invertedAxes)

Axis of second column is inverted.

cutoffs argument

histoVisibility <- rep(TRUE, ncol(iris))
cutoffs <- list(list(c(6, 7)), NULL, NULL, NULL, c("virginica", "setosa"))
parallelPlot(iris, histoVisibility = histoVisibility, cutoffs = cutoffs)

Lines which are not kept by cutoffs are shaded; an histogram is displayed considering only kept lines.

refRowIndex argument

parallelPlot(iris, refRowIndex = 1)

Axes are shifted vertically in such a way that first trace of the dataset looks horizontal.

rotateTitle argument

parallelPlot(iris, refColumnDim = "Species", rotateTitle = TRUE)

Column names are rotated (can be useful for long column names).

columnLabels argument

columnLabels <- gsub("\\.", "<br>", colnames(iris))
parallelPlot(iris, refColumnDim = "Species", columnLabels = columnLabels)

Given names are displayed in place of column names found in dataset; <br> is used to insert line breaks.

cssRules argument

parallelPlot(iris, cssRules = list(
    "svg" = "background: white", # Set background of plot to white
    ".axisLabel" = c("fill: red", "font-size: 1.8em"), # Set title of axes red and greater
    ".tick text" = "font-size: 1.8em", # Set text of axes ticks greater
    ".plotGroup path" = "opacity: 0.25", # Make lines less opaque
    ".xValue" = "color: orange", # Set color for x values in tooltip
    ".xName" = "display: table" # Trick to have x values on a new line
))

Apply CSS to the plot. CSS is a simple way to describe how elements on a web page should be displayed (position, color, size, etc.). You can learn the basics at W3Schools. You can learn how to examine and edit css at MDN Web Docs for Firefox or Chrome devtools for Chrome.

sliderPosition argument

parallelPlot(iris, sliderPosition = list(
  dimCount = 3, # Number of columns to show
  startingDimIndex = 2 # Index of first shown column
))
# Visible columns starts at second column and three columns are represented.

Set initial position of slider, specifying which columns interval is visible.

controlWidgets argument

parallelPlot(iris, refColumnDim = "Species", controlWidgets = TRUE)

Some widgets are available to control the plot.