grouped_stats: Grouped Weighted Means, Quartiles, or Variances

In flashlight: Shed Light on Black Box Machine Learning Models

Grouped Weighted Means, Quartiles, or Variances

Description

Calculates weighted means, quartiles, or variances (and counts) of a variable grouped by optional columns. By default, counts are not weighted, even if there is a weighting variable.

Usage

grouped_stats(
  data,
  x,
  w = NULL,
  by = NULL,
  stats = c("mean", "quartiles", "variance"),
  counts = TRUE,
  counts_weighted = FALSE,
  counts_name = "counts",
  value_name = x,
  q1_name = "q1",
  q3_name = "q3",
  ...
)

Arguments

data	A data.frame.
x	Variable name in data to summarize.
w	Optional name of the column in data with case weights.
by	An optional vector of column names in data used to group the results.
stats	Statistic to calculate: "mean", "quartiles", or "variance".
counts	Should group counts be added?
counts_weighted	Should counts be weighted by the case weights? If TRUE, the sum of w is returned by group.
counts_name	Name of column in the resulting data.frame containing the counts.
value_name	Name of the resulting column with mean, median, or variance.
q1_name	Name of the resulting column with first quartile values. Only relevant if stats = "quartiles".
q3_name	Name of the resulting column with third quartile values. Only relevant if stats = "quartiles".
...	Additional arguments passed to corresponding ⁠weighted_*()⁠ functions in MetricsWeighted.

Value

A data.frame with columns by, x, and optionally counts_name.

Examples

grouped_stats(iris, "Sepal.Width")
grouped_stats(iris, "Sepal.Width", stats = "quartiles")
grouped_stats(iris, "Sepal.Width", stats = "variance")
grouped_stats(iris, "Sepal.Width", w = "Petal.Width", counts_weighted = TRUE)
grouped_stats(iris, "Sepal.Width", by = "Species")

flashlight documentation built on May 31, 2023, 6:19 p.m.