Add individual data points and data labels to lattice boxplots

[Macondo_Cuanto]

I'm learning how to use the lattice package, and I'm not sure how to add individual data points to boxplots.

The next challenge would be to add measurement ID data labels to each point.

This is what desired output looks like in Minitab:

And this is the best I can do with lattice plots:

I can add individual points and labels using ggplot2, but I'd like to explore lattice package capabilities.

Would someone please help me out here?

Example:

#load libraries
library(lattice)
library(latticeExtra)

#make some data
my_data <- PlantGrowth
my_data$MeasurementID <- sprintf("M%s",seq(1:nrow(my_data)))

#plot those data
bwplot(weight ~ group,
       data = my_data,
       axis=axis.grid,
       main="Plant Growth Dataset (Plotted in R)",
       sub = "I would like to add individual data points for each boxplot \nAdding MeasurementID point labels would be a welcome bonus",
       scales = "free",
       xlab = "Treatment",
       ylab = "Plant Weight",
       par.settings = list(
         box.rectangle = list(
           fill= rep(c('#0071BF','#E9BB00','#838383'),3),
           col = "black"),
         box.umbrella = list(
           col = "black"),
         plot.symbol = list(
           col = "black",
           pch = 8),
         strip.background= list(
           col="lightgrey"),
         par.sub.text = list(
           fontsize = 10,
           font = 1,
           just= "left",
           x = grid::unit(5, "mm")),
         par.main.text = list(
           fontsize = 18)
       ))

#add individual data points to each boxplot
???

#label those data points with values in the MeasurementID field
???

[jrkrideau]

I am not a lattice user but I think as.layer in {latticeExtra} may give you the scatter-plot. For the labels, have a look at How to label points on a scatterplot with R (for lattice)

[Macondo_Cuanto]

Thanks for the suggestion, @jrkrideau. You were right about the as.layer function.

Using the resources you directed me to, I found 2 solutions to my problem.

1. Using lattice and latticeExtra packages
2. Using a third package, latticetools, to add optional callout lines between labels and points

Apologies for the odyssey that follows; it's not Shakespeare but it's serviceable.

Solution 1:

# load libraries
library(lattice)
library(latticeExtra)

# make some data
my_data <- PlantGrowth
my_data$MeasurementID <- sprintf("M%s", seq(1:nrow(my_data)))

# plot those data
bwplot(weight ~ group,
  data = my_data,
  axis = axis.grid,
  main = "Plotting individual point symbols and labels using R",
  sub = "Uses lattice & latticeExtra packages",
  scales = "free",
  xlab = "Treatment",
  ylab = "Plant Weight",
  # use line instead of dot for median symbol
  pch = "|",
  # additional boxplot element settings
  par.settings = list(
    box.rectangle = list(
      # use these box fill colors
      fill = rep(c("#3985E4", "#FFEC48", "#999999"), 3),
      # make box outlines black
      col = "black"
    ),
    # make whiskers black
    box.umbrella = list(
      col = "black"
    ),
    # make outlier symbols red stars
    plot.symbol = list(
      col = "red",
      pch = 8
    ),
    # make title text 18pt
    par.main.text = list(
      fontsize = 14
    )
  )
) +
  # add individual points to existing plot
  as.layer(xyplot(weight ~ group,
    # select dataframe
    data = my_data,
    # function that adds labels
    panel = function(x, y, ...) {
      panel.xyplot(x, y, ...)
      panel.text(x, y,
        # select field that has point labels
        labels = my_data$MeasurementID,
        # make labels black
        col = "black",
        # set label size
        cex = 0.5,
        # set label position; 4 is to the right of points
        pos = 4,
        # this controls how far labels are drawn from points
        offset = 1
      )
    },
    # tweak individual point display
    par.settings = list(
      plot.symbol = list(
        # make individual points black
        col = "black",
        # make individual points open circles
        pch = 1
      )
    )
  ))

Solution 2:

# load libraries
library(lattice)
library(latticeExtra)
library(latticetools)

# make some data
my_data <- PlantGrowth
my_data$MeasurementID <- sprintf("M%s", seq(1:nrow(my_data)))

# plot the data
bwplot(weight ~ group,
  data = my_data,
  axis = axis.grid,
  main = "Plotting individual point symbols and labels using R",
  sub = "Uses latticetools package. Install with:\n devtools::install_github('https://github.com/m-jahn/lattice-tools')",
  scales = "free",
  xlab = "Treatment",
  ylab = "Plant Weight",
  # use line instead of dot for median symbol
  pch = "|",
  # additional boxplot element settings
  par.settings = list(
    box.rectangle = list(
      # use these box fill colors
      fill = rep(c("#3985E4", "#FFEC48", "#999999"), 3),
      # make box outlines black
      col = "black"
    ),
    # make whiskers black
    box.umbrella = list(
      col = "black"
    ),
    # make outlier symbols red stars
    plot.symbol = list(
      col = "red",
      pch = 8
    ),
    # make title text 18pt
    par.main.text = list(
      fontsize = 14
    )
  )
) +
  # add individual points
  as.layer(xyplot(weight ~ group,
    # select dataframe
    data = my_data,
    # select field that has point labels
    labels = my_data$MeasurementID,
    # set label size
    cex = 0.5,
    # function that actually adds labels
    panel = function(x, y, ...) {
      panel.xyplot(x, y, ...)
      panel.directlabels(x, y, ...)
    },
    # tweak individual point display
    par.settings = list(
      plot.symbol = list(
        # make individual points black
        col = "black",
        # make individual points open circles
        pch = 1
      )
    )
  ))

[jrkrideau]

Lovely. Better than I had hope for.