autoplot.funData: Visualize functional data objects using ggplot

Description

This function allows to plot funData objects based on the ggplot2 package. The function provides a wrapper that rearranges the data in a funData object on a one- or two-dimensional domain and provides a basic ggplot object, which can be customized using all functionalities of the ggplot2 package.

Usage

autoplot.funData(
  object,
  obs = seq_len(nObs(object)),
  geom = "line",
  plotNA = FALSE,
  ...
)
autolayer.funData(
  object,
  obs = seq_len(nObs(object)),
  geom = "line",
  plotNA = FALSE,
  ...
)

Value

A ggplot object that can be customized using all functionalities of the ggplot2 package.

Arguments

object: A funData object on a one- or two-dimensional domain.
obs: A vector of numerics giving the observations to plot. Defaults to all observations in object. For two-dimensional functions (images) obs must have length 1.
geom: A character string describing the geometric object to use. Defaults to "line". See ggplot2 for details.
plotNA: Logical. If TRUE, missing values are interpolated using the approxNA function (only for one-dimensional functions). Defaults to FALSE. See Details.
...: Further parameters passed to geom_line (for one dimensional domains, e.g. alpha, color, fill, linetype, size) or to geom_raster (for two-dimensional domains, e.g. hjust, vjust, interpolate).

Details

If some observations contain missing values (coded via NA), the functions can be interpolated using the option plotNA = TRUE. This option relies on the na.approx function in package zoo and is currently implemented for one-dimensional functions only in the function approxNA.

Examples

Run this code

# Install / load package ggplot2 before running the examples
library("ggplot2")

# One-dimensional
argvals <- seq(0,2*pi,0.01)
object <- funData(argvals,
                   outer(seq(0.75, 1.25, length.out = 11), sin(argvals)))

g <- autoplot(object) # returns ggplot object
g # plot the object

# add the mean function in red
g + autolayer(meanFunction(object),  col = 2)

# Two-dimensional
X <- array(0, dim = c(2, length(argvals), length(argvals)))
X[1,,] <- outer(argvals, argvals, function(x,y){sin((x-pi)^2 + (y-pi)^2)})
X[2,,] <- outer(argvals, argvals, function(x,y){sin(2*x*pi) * cos(2*y*pi)})
object2D <- funData(list(argvals, argvals), X)

# \donttest{
autoplot(object2D, obs = 1)
autoplot(object2D, obs = 2)# }
if (FALSE) autoplot(object2D) # must specify obs!

### More examples ###
# \donttest{
par(mfrow = c(1,1))

# using plotNA (needs packages zoo and gridExtra)
requireNamespace("zoo", quietly = TRUE)
requireNamespace("gridExtra", quietly = TRUE)
objectMissing <- funData(1:5, rbind(c(1, NA, 5, 4, 3), c(10, 9, NA, NA, 6)))
g1 <- autoplot(objectMissing) # the default
g2 <- autoplot(objectMissing, plotNA = TRUE) # requires zoo

gridExtra::grid.arrange(g1 + ggtitle("plotNA = FALSE (default)"),
                        g2 + ggtitle("plotNA = TRUE")) # requires gridExtra

# Customizing plots (see ggplot2 documentation for more details)
# parameters passed to geom_line are passed via the ... argument
gFancy <- autoplot(object, color = "red", linetype = 2) 
gFancy

# new layers can be added directly to the ggplot object
gFancy + theme_bw() # add new layers to the ggplot object
gFancy + ggtitle("Fancy Plot with Title and Axis Legends") + 
         xlab("The x-Axis") + ylab("The y-Axis")

autoplot(object2D, obs = 1) + ggtitle("Customized 2D plot") + theme_minimal() +
          scale_fill_gradient(high = "green", low = "blue", name = "Legend here")
# }