ggplot2

Use special symbol in labels

• Text labels: parse = T inside geom_text() or annotate()
• Axis labels: expression(alpha) to get greek alpha
• Facet labels: Use labeller = label_parsed() inside facet
• Legend labels: Use bquote(alpha == .(value))

Note:

• bold('bold'), italic('italic'), atop('top','bottom')
• * is a non space separator, ~ is to add spaces
• [] for subscript and ^ for superscript
• bquote() when substitutions are needed .(sub) to substitute an object
• help('plotmath') for more info

Example: labs(x = expression(bold(T[max])~(h))

Transparent rectangles

Add transparent rectangle to highlight graph area.

# Create data for rectangle
rect <- data.frame(xmin = 2, xmax = 4, ymin = -Inf, ymax = Inf)

# Create plot
ggplot(data = mtcars, aes(x = wt, y = hp)) +
  geom_point() +
  geom_rect(data = rect, aes(xmin = xmin, xmax = xmax, 
                             ymin = ymin, ymax = ymax),
            fill = 'coral', alpha = 0.4, inherit.aes = FALSE)

Add mean or median to dotplot

Not working with ggplot2 v2.0.1

ggplot(data = mtcars, aes(x = cyl, y = hp, color = cyl)) + 
  geom_point() +
  geom_errorbar(stat = 'hline', yintercept = 'mean', width = 0.8, 
                aes(ymax = ..y.., ymin = ..y..))

Scale geom_density() to geom_histogram().

ggplot(mtcars, aes(x = hp)) +
  geom_histogram() +
  geom_density(aes(x = hp, 4 * ..scaled..), inherit.aes = FALSE)

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Correlation graphs

Uses the package GGally.

library(GGally)
ggpairs(data, columns = 1:7, colour = 'sex', 
        lower = list(continuous = 'smooth'), 
        diag  = list(continuous = 'density', discrete = 'bar'),
        axisLabels = 'show')

gridExtra

The function grid.arrange() draws directly on a device while arrangeGrob() doesn't draw anything but returns a grob.

# Plot 4 graphs with different panel size
grid.arrange(P1, P2, P3, P4, 
             ncol = 2, nrow = 2, 
             widths = c(4, 1), heights = c(1, 4))

Work around smooth curves

Classic smooth

ggplot(iris, aes(x = Sepal.Length, y = Sepal.Width)) +
  geom_point() +
  geom_smooth()

Special smooth

The special smooth visually gives less importance to edges of the smooth curve by plotting data outside 90% confidence interval in light grey.

# 90% CI of the Independent variable
CI90     <- quantile(iris$Sepal.Length, c(0.05, 0.95))

# Generate ggplot with smooth
Gsmooth  <- ggplot(data = iris, 
                   aes(x = Sepal.Length, y = Sepal.Width)) +
  stat_smooth(method = 'loess', se = FALSE)

# Get the data from the smooth
Gsmooth  <- as.data.frame(ggplot_build(Gsmooth)$data)

# Subset the 90% CI of the smooth
Gsmooth  <- Gsmooth[Gsmooth$x >= CI90[1] & Gsmooth$x <= CI90[2], ]

p <- ggplot()

# Rectangle showing the 90%CI
p <- p + geom_rect(aes(xmin = CI90[1], xmax = CI90[2],
                       ymax = Inf, ymin = -Inf), 
                   alpha = 0.2, fill = 'dodgerblue3')

# Full smooth as a grey line
p <- p + geom_smooth(data = iris, aes(x = Sepal.Length, y = Sepal.Width), 
                     method = 'loess', 
                     se = FALSE, 
                     size = 1, 
                     col = 'grey70')

# 90% CI smooth as black continuous line
p <- p + geom_line(data = Gsmooth, aes(x = x, y = y), size = 1, col = 'black')

# Add actual data
p <- p + geom_point(data = iris, aes(x = Sepal.Length, y = Sepal.Width), 
                    col = 'grey35')

  # Add labels
p <- p + annotate('text', x = mean(CI90), y = -Inf, 
                  label = '90% confidence interval',
                  vjust = -0.5,
                  col = 'dodgerblue4', alpha = 0.7, size = 4)
print(p)

Scales

New color schemes (+ invert color scale)

scale_fill_manual(values = rev(RColorBrewer::brewer.pal(6, 'Reds')))

For more information visit the ColorBrewer page

Remove legend for a given aesthetic

scale_linetype_identity()

Remove an axis

scale_x_continuous(NULL, breaks = NULL)

Expand the limits

expand_limits()

Special breaks for each panel

ebreaks <- function(...) {
  function(x) {    # x contains min and max for each panel
    if (x[1] < 120) { 
      breaks <- seq(0, 200, 10)
    } else { 
      breaks <- seq(150, 350, 50)
    }
    return(breaks) # Return all breaks
  }}

ggplot(mtcars, aes(x = hp, y = disp)) +
  geom_point() +
  scale_x_continuous(breaks = ebreaks()) +
  facet_wrap(~ cyl, scales = 'free_x')

Annotations

Based on graph coordinates

annotate(..., nudge_y = 3, nudge_x = 10)

Based on relative positioning

annotate(..., hjust = 1 , vjust = 0)

Special annotations

annotation_custom()    # Special annotation function
annotation_logticks()  # For creating nice logscale

Control the legend

guides(fill = guide_legend(reverse = TRUE))

See also guide_colorbars()

Access the ggplot2 lm() data

fortify.lm()

Use the labeller function

ggplot(mtcars, aes(x = disp, y = mpg)) + 
  geom_point() + 
  facet_wrap(~ vs + am, 
             label = labeller(.default = label_both, .multi_line = FALSE))

Sliders with manipulate

library(manipulate)
manipulate(
  ggplot(data = mtcars, aes(x = hp , y = cyl)) + 
    geom_line(), 
  prm = slider(0, 20, 10, step = 1)
)