autocol {paletteknife}

Map Colours From Value

Description

Create a vector of colours and associated legend for easier base plots

Usage

autocol(
  x,
  set = "",
  alpha = NA,
  limits = NA,
  na_colour = NA,
  bias = 1,
  legend_len = 6
)

palette.misc

palette.viridis

palette.colorbrewer

Arguments

x

Vector to be mapped to colours

set

Colour set to use – see Details for full list. A default sasha or viridis is chosen if empty.

alpha

Transparency as a single value or as another vector (recycled to fill). If it is a vector, all values are scaled from 0:max(alpha) meaning transparent:opaque. Single values must be in range 0-1. If NA no alpha channel is added.

limits

Colour scale limits as absolute range c(0,10) or NA = full range

na_colour

Colour to represent NA-values, default NA returns a colour of NA (thus not plotted)

bias

Skew to apply to colour-ramp (>1 increases resolution at low end, <1 at the high end)

legend_len

Continuous legend target size

Format

An object of class list of length 1.

An object of class list of length 8.

An object of class list of length 35.

Details

Helper function for using colours in R's default plot() and legend(). Colours from built-in palettes are automatically scaled to return a vector of colours and create options('autolegend') which contains the correct legend mapping for autolegend().

A discrete palette is used for factor and character inputs whilst a continuous palette is used for integer and numeric.

Colour sets built-in so far are held in lists starting pals. and can be visualized most easily with pals_display(). The set argument can be any of the colour set names listed here (such as 'magma'), or from palette.pals(), or finally as a custom-defined vector, such as set = rainbow(5).

The current lists of palettes included with paletteknife all being with pal.

• pals.viridis

  All of the continuous palette forked from the viridisLite package maintained by Simon Garnier.

  • Contains: cividis inferno magma mako plasma rocket turbo viridis

• pals.rcolorbrewer

  All of the palettes included in RColorBrewer

  • Categorical: Accent Set1 Set2 Set3 Paired Pastel1 Pastel2 Dark2

  • Continuous: Greys Blues BuGn BuPu Greens GnBu PuBu Purples PuBuGn YlGnBu YlGn YlOrBr YlOrRd Oranges OrRd Reds RdPu PuRd

  • Divergent: Spectral RdYlBu RdYlGn BrBG RdBu RdGy PiYG PRGn PuOr

• pals.misc

  • Sasha Trubetskoy (2017): List of 20 Simple, Distinct Colors: sasha

Custom limits can be specified using c(0,10). This is useful if multiple plots using the same range are required for cross-comparison. Default behaviour (limits = NA) sets the range to exactly fit.

The skew of the colourscale can be adjusted with bias, for example if x has an exponential distribution, a bias value > 1 will bring out contrast at the low end.

Value

A character vector of colours of equal length to input x, sampled from the chosen set. This allows it to be used for plotting directly. Information for a legend (containing every level for categorical data, or approximately length legend_len for continuous) is stored in options('autolegend') and not returned explicitly.

Examples

plot(iris$Sepal.Length, iris$Petal.Length, cex=3, pch=16,
    col=autocol(iris$Petal.Width, set='PuBuGn', alpha=0.8, legend_len=12) )
  autolegend('topleft', title='Petal.Width', ncol=3)
  # Also try simplest "autolegend()" for click-to-draw

# Try scales which include NA in both colour and alpha channel
with(airquality, plot(Temp, col=autocol(x=Solar.R, set='YlOrRd', alpha=Ozone,
    na_colour='cyan'), pch=16, cex=sqrt(Wind) ))
  # Note inset=1 draws on opposite side ie above not below plot area
  autolegend('bottom', inset=1, bty='n', horiz=TRUE)


# Here we want a summary plot ordered by level, so need to create a colour vector to match
# 'Alphabet' is a built-in colour set, see "palette.pals()"
mixedbag = as.factor(sample(letters,1000,replace=TRUE))
  plot(x=mixedbag, y=rnorm(1000), col=autocol(levels(mixedbag), set='Alphabet'))
  autolegend('bottom', ncol=9, cex=0.7)

# Maintain the order of strings
barplot(1:8, col=autocol(LETTERS[8:1]))
  autolegend('topleft')

# Any unusual formats are coerced to numeric and the legend converted back
mydates = as.Date('2000-01-01')+0:100
  plot(mydates, pch=16, col=autocol(mydates, set=rainbow(10), bias=2) )
  autolegend(x=0, y=mydates[100], title='My Dates')

# Timeseries objects plot as a line, but can overlay with points()
plot(airmiles)
  points(airmiles, pch=15, col=autocol(airmiles, set='Reds'))

# Use the limits to clip or augment the colour-scale
layout(matrix(1:2))
  plot(runif(10), col=autocol(1:10, limits=c(0,20)), pch=16,
    main='Data split over two plots with same scale')
  plot(runif(10), col=autocol(c(100,20:12), limits=c(0,20)), pch=16)
  text(1, 0.5, pos=4, xpd=NA,
'This point has a
value of 100 but
clipped to max
colour == 20')
  autolegend('bottom', inset=1, horiz=TRUE) # Draws above!
  layout(1)