create_qgroups: Create Quantile Groups

In dvmisc: Convenience Functions, Moving Window Statistics, and Graphics

Description

Combines quantile and cut into a single function, with strata-specific quantiles possible. For example, you could create sex-specific height tertiles with create_qgroups(height, groups = 3, strata = sex). Compatible with dplyr functions like mutate and transmute.

Usage

create_qgroups(x, groups = 4, probs = seq(1/groups, 1 - 1/groups,
  1/groups), strata = NULL, quantile_list = list(na.rm = TRUE),
  cut_list = list(include.lowest = TRUE))

Arguments

x	Numeric vector.
groups	Numeric value, e.g. 3 for tertiles, 4 for quartiles, etc.
probs	Numeric vector.
strata	Factor specifying subgroups to calculate quantiles within. For multivariable subgroups, you can use interaction.
quantile_list	Arguments to pass to quantile.
cut_list	Arguments to pass to cut.

Value

Factor variable.

Examples

# In mtcars dataset, create tertiles for mpg
mtcars$mpg_tertiles <- create_qgroups(mtcars$mpg, groups = 3)
table(mtcars$mpg_tertiles)

# Define tertile cutpoints separately for 4-, 6-, and 8-cylinder vehicles
mtcars$mpg_tertiles <- create_qgroups(mtcars$mpg, groups = 3, strata = mtcars$cyl)
table(mtcars$mpg_tertiles)

# Works with dplyr functions like mutate
mtcars <- mtcars %>% 
  dplyr::mutate(mpg_tertiles = create_qgroups(mpg, groups = 3, strata = cyl))
table(mtcars$mpg_tertiles)

# Can embed in lm, glm, etc.
summary(lm(mpg ~ create_qgroups(wt), data = mtcars))

dvmisc documentation built on Dec. 18, 2019, 1:35 a.m.