Create a survey data transformation recipe

Creates a Recipe object that encapsulates a sequence of data transformations that can be applied to surveys in a reproducible manner. Recipes allow documenting, sharing, and reusing data processing workflows.

Usage

recipe(...)

Arguments

...

Required metadata and optional steps. Required parameters:

• name: Descriptive name for the recipe

• user: User/author creating the recipe

• svy: Base Survey object (use survey_empty() for generic recipes)

• description: Detailed description of the recipe's purpose

Optional parameters include data transformation steps.

Value

A Recipe object containing metadata, transformation steps, dependency information, and default engine configuration.

Details

Recipes are essential for:

• Reproducibility: Ensure transformations are applied consistently

• Documentation: Keep a record of what transformations are performed and why

• Collaboration: Share workflows between users and teams

• Versioning: Maintain different processing versions for different editions

• Automation: Apply complex transformations automatically

Steps included in the recipe can be any combination of step_compute, step_recode, or other transformation steps.

Recipes can be saved with save_recipe(), loaded with read_recipe(), and applied automatically with bake_recipes().

See also

Recipe for class definition save_recipe to save recipes read_recipe to load recipes get_recipe to retrieve recipes from repository bake_recipes to apply recipes to data

Other recipes: Recipe-class, add_recipe(), bake_recipes(), explore_recipes(), get_recipe(), print.Recipe(), publish_recipe(), read_recipe(), save_recipe(), steps_to_recipe()

Examples

# Basic recipe without steps
r <- recipe(
  name = "Basic ECH Indicators",
  user = "Analyst",
  svy = survey_empty(type = "ech", edition = "2023"),
  description = "Basic labor indicators for ECH 2023"
)
r
#> 
#> ── Recipe: Basic ECH Indicators ──
#> Author:  Analyst
#> Survey:  ech / 2023
#> Version: 1.0.0
#> Description: Basic labor indicators for ECH 2023
#> Certification: community
#> 

# \donttest{
# Recipe with steps using local data
dt <- data.table::data.table(
  id = 1:50, age = sample(18:65, 50, TRUE),
  income = runif(50, 1000, 5000), w = runif(50, 0.5, 2)
)
svy <- Survey$new(
  data = dt, edition = "2023", type = "demo",
  psu = NULL, engine = "data.table",
  weight = add_weight(annual = "w")
)
svy <- svy |>
  step_compute(income_cat = ifelse(income > 3000, "high", "low")) |>
  step_recode(age_group, age < 30 ~ "young", .default = "adult")
r2 <- recipe(
  name = "Demo", user = "test", svy = svy,
  description = "Demo recipe", steps = get_steps(svy)
)
r2
#> 
#> ── Recipe: Demo ──
#> Author:  test
#> Survey:  demo / 2023
#> Version: 1.0.0
#> Description: Demo recipe
#> Certification: community
#> 
#> ── Requires (2 variables) ──
#>   income, age
#> 
#> ── Pipeline (2 steps) ──
#>   1. [compute] -> income_cat  "Compute step"
#>   2. [recode] -> age_group  "Recode step"
#> 
#> ── Produces (2 variables) ──
#>   age_group [categorical], income_cat [numeric]
#> 
# }