box.dt: Box Plot as data.table

In VEuPathDB/plot.data: Create `data.table`s of Ready to Plot Data

Box Plot as data.table

Description

This function returns a data.table of plot-ready data with one row per group (per panel). Columns 'x', 'min', 'q1', 'median', 'q3' and 'max' represent the pre-computed values per group. Columns 'group' and 'panel' specify the group the data belong to. Optionally, can return columns 'outliers' and 'mean' as well.

Usage

box.dt(
  data,
  variables,
  points = c("outliers", "all", "none"),
  mean = c(FALSE, TRUE),
  computeStats = c(FALSE, TRUE),
  overlayValues = NULL,
  sampleSizes = c(TRUE, FALSE),
  completeCases = c(TRUE, FALSE),
  evilMode = c("noVariables", "allVariables", "strataVariables"),
  verbose = c(TRUE, FALSE)
)

Arguments

data	data.frame to make plot-ready data for
variables	veupathUtils VariableMetadataList
points	character vector indicating which points to return 'outliers' or 'all'
mean	boolean indicating whether to return mean value per group (per panel)
computeStats	boolean indicating whether to compute nonparametric statistical tests (across x values or group values per panel)
overlayValues	veupathUtils::BinList providing overlay values of interest
sampleSizes	boolean indicating if sample sizes should be computed
completeCases	boolean indicating if complete cases should be computed
evilMode	String indicating how evil this plot is ('strataVariables', 'allVariables', 'noVariables') Metadata can include 'displayName', 'displayRangeMin', 'displayRangeMax', and 'collectionVariable'. Will be included as an attribute of the returned plot object.
verbose	boolean indicating if timed logging is desired

Value

data.table plot-ready data

Evil Mode

An 'evilMode' exists. It will do the following:
- when ‘strataVariables' it will return ’no data' as a regular value for strata vars but will discard such cases for the axes vars.
- when ‘allVariables' it will return ’no data' as a regular value for all variables.
- when 'noVariables' it will do the sensible thing and return complete cases only.
- not return statsTables
- allow smoothed means and agg values etc over axes values where we have no data for the strata vars
- return a total count of plotted incomplete cases
- represent missingness poorly, conflate the stories of completeness and missingness, mislead you and steal your soul

Examples

# Construct example data
df <- data.table('entity.xvar' = sample(c('a','b','c'), 100, replace=T),
                 'entity.yvar' = rnorm(100),
                 'entity.overlay' = sample(c('red','green','blue'), 100, replace=T))

# Create VariableMetadataList that specifies variable role in the plot and supplies variable metadata
variables <- veupathUtils::VariableMetadataList(
  veupathUtils::VariableMetadata(
    variableClass = veupathUtils::VariableClass(value = 'native'),
    variableSpec = veupathUtils::VariableSpec(variableId = 'xvar', entityId = 'entity'),
    plotReference = veupathUtils::PlotReference(value = 'xAxis'),
    dataType = veupathUtils::DataType(value = 'STRING'),
    dataShape = veupathUtils::DataShape(value = 'CATEGORICAL')
  ),
  veupathUtils::VariableMetadata(
    variableClass = veupathUtils::VariableClass(value = 'native'),
    variableSpec = veupathUtils::VariableSpec(variableId = 'overlay', entityId = 'entity'),
    plotReference = veupathUtils::PlotReference(value = 'overlay'),
    dataType = veupathUtils::DataType(value = 'STRING'),
    dataShape = veupathUtils::DataShape(value = 'CATEGORICAL')
  ),
  veupathUtils::VariableMetadata(
    variableClass = veupathUtils::VariableClass(value = 'native'),
    variableSpec = veupathUtils::VariableSpec(variableId = 'yvar', entityId = 'entity'),
    plotReference = veupathUtils::PlotReference(value = 'yAxis'),
    dataType = veupathUtils::DataType(value = 'NUMBER'),
    dataShape = veupathUtils::DataShape(value = 'CONTINUOUS')
  )
)

# Returns a data table with plot-ready data
dt <- box.dt(df, map, points = 'outliers', mean=F, computeStats=T)

VEuPathDB/plot.data documentation built on Feb. 20, 2025, 6:33 p.m.