superbPlot 'corset' layout

Source: R/functionsPlotting_corset.R

superbPlot comes with a few built-in templates for making the final plots. The corset plot is specifically devised for 2-repeated-measure design: it merges the "pointindividualline" layout with a raincloud layout (Belisario 2021) . All layout produces ggplot objects that can be further customized. Additionally, it is possible to create custom-make templates (see vignette 5). The functions, to be "superbPlot-compatible", must have these parameters:

superbPlot.corset(
  summarydata,
  xfactor,
  groupingfactor,
  addfactors,
  rawdata = NULL,
  lineParams = list(),
  pointParams = list(),
  errorbarParams = list(),
  jitterParams = list(),
  violinParams = list(),
  facetParams = list(),
  xAsFactor = TRUE
)

Arguments

summarydata

a data.frame with columns "center", "lowerwidth" and "upperwidth" for each level of the factors;

xfactor

a string with the name of the column where the factor going on the horizontal axis is given;

groupingfactor

a string with the name of the column for which the data will be grouped on the plot;

addfactors

a string with up to two additional factors to make the rows and columns panels, in the form "fact1 ~ fact2";

rawdata

always contains "DV" for each participants and each level of the factors;

lineParams

(optional) list of graphic directives that are sent to the geom_line layer including the option colorize "bySlope", "byId" or "none";

pointParams

(optional) list of graphic directives that are sent to the geom_bar layer;

errorbarParams

(optional) list of graphic directives that are sent to the geom_superberrorbar layer;

jitterParams

(optional) list of graphic directives that are sent to the geom_jitter layer;

violinParams

(optional) list of graphic directives that are sent to the geom_boxplot layer;

facetParams

(optional) list of graphic directives that are sent to the facet_grid layer;

xAsFactor

(optional) Boolean to indicate if the factor on the horizontal should continuous or discrete (default is discrete).

Value

a ggplot object

References

Belisario K (2021). Corset Plots: Visualizing Heterogeneity in Change Outcomes Across Two Timepoints. doi:10.5281/zenodo.4905031 , https://cran.r-project.org/package=ggcorset.

Examples

# We first generate randomly a 2-measurement dataset with 50 participants and a large effect
dta <- GRD(SubjectsPerGroup = 50, WSFactors = "moment(2)", Effects = list("moment"=slope(3)))

# This will make a basic corset plot 
superb(
   cbind(DV.1, DV.2) ~ .,
   dta, 
   WSFactors = "moment(2)", 
   plotLayout = "corset" 
)


# This will color the increasing and decreasing individuals
superb(
   cbind(DV.1, DV.2) ~ ., 
   dta, 
   WSFactors = "moment(2)", 
   plotLayout = "corset",
   lineParams = list(colorize="bySlope") 
)


# This layout has similarities with the "pointindividualline" layout
superb(
   cbind(DV.1, DV.2) ~ ., 
   dta, 
   WSFactors = "moment(2)", 
   plotLayout = "pointindividualline" 
)


# if you extract the data with superbData, you can 
# run this layout directly
processedData <- superb(
   cbind(DV.1, DV.2) ~ ., 
   dta, 
   WSFactors = "moment(2)", 
   showPlot  = FALSE
)

superbPlot.corset(processedData$summaryStatistic, 
   "moment", NULL, ".~.", 
   processedData$rawData, 
   lineParams = list(colorize="bySlope") )