Create Dummy Variables (One-Hot Encoding)

Creates 0/1 dummy variables from categorical variables. Column names are derived from value labels when available, making results readable for SPSS-style data.

Unlike model.matrix(), to_dummy() uses value labels for column naming and handles haven_labelled vectors correctly.

Usage

to_dummy(data, ..., suffix = "val", ref = NULL, append = TRUE)

Arguments

data: A data frame or vector. When a data frame is passed, use ... to select variables.
...: Variables to dummy-code (tidyselect). Only used when data is a data frame.
suffix: How to name dummy columns: "val" (default) uses the raw value (e.g., gender_1), "label" uses the value label (e.g., gender_Male).
ref: A value to use as reference category (omitted from output). If NULL (default), all categories get a dummy variable. Set to a specific value for n-1 coding (e.g., for regression). For a factor, give a level name (ref = "Male") or a number: the original code of a to_label() factor, otherwise the level position (ref = 1 = first level). A ref that matches no category is an error.
append: If TRUE (default), the dummy columns are appended to the original data frame. If FALSE, only the dummy columns are returned. Ignored when data is a vector.

Value

If append = TRUE (default), the original data frame with dummy columns appended. If append = FALSE, a tibble with only the dummy columns. For vector input, always a tibble of dummy columns.

Details

Column Naming

With suffix = "val": {varname}_{value} (e.g., gender_1, gender_2). With suffix = "label": {varname}_{label} where labels are cleaned (umlauts transliterated, e.g. "männlich" -> maennlich; spaces replaced with _; other special characters removed). Values without a usable label (e.g. "..") use their value; labels shared by several values get the value appended, so every category keeps its own column.

Reference Category

For regression, you typically need n-1 dummy variables. Set ref to the value of the reference category to omit it.

Examples

library(dplyr)
data(survey_data)

# Create dummies and append to data (default)
data <- to_dummy(survey_data, gender)

# Use labels for column names
data <- to_dummy(survey_data, gender, suffix = "label")

# n-1 dummies with reference category
data <- to_dummy(survey_data, gender, ref = 1)

# Multiple variables
data <- to_dummy(survey_data, gender, education, suffix = "label")

# Return only the dummy columns (without original data)
dummies <- to_dummy(survey_data, gender, append = FALSE)