aurora

Descriptive Statistics Graphics and Utilities

Version 0.0.0.927 · License GPL (≥ 2)

Overview

aurora is the graphics package of the DescToolsX ecosystem. It provides a comprehensive collection of statistical plots, geometric drawing primitives, color tools, annotation helpers, and formatting utilities — all built directly on base R graphics.

Building on base graphics keeps the package lightweight and fully compatible with everything base R offers (layout(), par(), custom devices), while a central theme system and consistent argument conventions remove the usual boilerplate: sensible defaults, automatic margin handling, and uniform styling across all plot functions.

aurora is part of the DescToolsX package suite:

Package	Role
bedrock	core infrastructure and utilities
aurora	graphics and visualization (this package)
lumen	statistical tests and confidence intervals
alloy	model fitting and evaluation
hermes	reporting and output

Installation

# development version from GitHub
remotes::install_github("AndriSignorell/aurora")

Key Features

Statistical plots

A large family of high-level plot*() functions with a consistent interface (formula support, grouped variants, theme-driven styling):

• Distributions: plotDens(), plotDens2D(), plotDensBox(), plotViolin(), plotRidge(), plotBox(), plotECDF(), plotFdist(), plotQQ(), plotProbDist()
• Categorical data: plotBar(), plotDot(), plotMosaic(), plotCatDist(), plotTreemap(), plotWeb()
• Relationships: plotXY(), plotLines(), plotCor(), plotAssoc(), plotBubble(), plotHexbin(), plotBag()
• Special purpose: plotTimeSeries(), plotArea(), plotMiss(), plotPropCI(), plotCirc(), plotPolar(), plotTernary(), plotBinaryTree(), plotFun()

Plot methods for objects from the suite are included, e.g. plot.Desc.* (for desc() results), plot.Lc (Lorenz curves), plot.BlandAltman, and lines() methods for lm and loess fits.

Theme system

All plot functions draw their defaults from a central, replaceable theme:

getTheme()                              # inspect the active theme
setTheme(list(palette = "Set2"))        # change one component globally
setTheme(list(grid = list(col = "grey90", lwd = 1, lty = "dotted")))
resetTheme()                            # back to the default

Explicit arguments always override the theme at the call site; the theme in turn overrides the package defaults. One place to define the look — every plot follows.

Geometric drawing

Geometry follows a clean two-step design: constructors create geometry objects — circle(), ellipse(), arc(), bezier(), band(), ring(), regPolygon() — and an overloaded polygon() generic (fully compatible with graphics::polygon()) draws them. canvas() provides a blank, aspect-true plotting canvas, polarGrid() a polar coordinate system.

canvas()
polygon(circle(radius = 1), col = "lightblue")
polygon(regPolygon(radius = 0.7, numVertices = 5), border = "red")

Because geometries are plain objects, they can be transformed before drawing (rotate(), transformXY()) or combined into composite shapes.

Colors

• Manipulation: addAlpha(), fade(), darken(), lighten(), mixColors(), grayscale(), colToOpaque()
• Analysis: contrastColor() (legible text colors on any background), findColor() (nearest named color)
• Conversion: hex, RGB, HSV, CMY(K), and long integer representations (colToHex(), hexToRGB(), cmykToRgb(), rgbToLong(), …)
• Palettes: pal() and palNames() for the suite's curated palettes

Annotation and layout helpers

Utilities that handle the fiddly parts of base graphics:

• stamp() — automatic plot stamping (author/date), theme-controlled
• boxedText(), barText(), textLegend(), colLegend() — labels and legends beyond text()/legend()
• errBars(), shade(), splineCI(), band() — uncertainty display
• axisBreak(), axTicks(), titleRect(), lineSep() — axis and title furniture
• spreadOut() — de-overlapping label positions
• isValidPlotRegion() — check the device geometry before drawing

Formatting and output

• fm() / fmCI() — flexible number and confidence interval formatting
• notation(), style() — notation and style control
• as.html(), toHtmlTable(), preview() — HTML rendering of tables and plots, e.g. for quick reports

Coordinates, units, and strings

• Coordinate transformations: transformXY(), rotate(), degToRad(), lineToUser(), abcCoords()
• Unit conversion engine: convUnit() with SI and derived units
• A complete str*() family for string handling (strTrim(), strPad(), strAlign(), strExtract(), strDist(), strAbbr(), …)

Example

library(aurora)

# grouped violin plot, styled by the active theme
plotViolin(mpg ~ cyl, data = mtcars)

# correlation matrix plot
plotCor(cor(mtcars))

# custom geometric graphic
canvas(main = "aurora primitives")
polygon(circle(radius = 1), col = addAlpha("steelblue", 0.4))
polygon(regPolygon(radius = 0.7, numVertices = 6), border = "red")

Design principles

• Base graphics, no grid — lightweight, transparent, hackable
• Consistent API — lowerCamelCase, uniform argument names and ordering across the whole suite
• Theme-driven — one place to define the look, every plot follows
• Robust by default — automatic margin adjustment, device geometry checks, protected graphics state
• Performance-aware — Rcpp under the hood where it matters

Documentation

• Website: https://andrisignorell.github.io/aurora/
• Bug reports: https://github.com/AndriSignorell/aurora/issues

License

GPL (≥ 2)