beezdemand (version 0.2.0)

pivot_demand_data: Reshape Demand Data Between Wide and Long Formats

Description

Converts demand data between wide format (one row per subject, prices as columns) and long format (one row per observation with id, x, y columns). This is a convenience wrapper around tidyr::pivot_longer() and tidyr::pivot_wider() tailored for behavioral economic purchase task data.

Usage

pivot_demand_data(
  data,
  format = c("long", "wide"),
  id_var = "id",
  x_var = "x",
  y_var = "y",
  price_cols = NULL,
  x_values = NULL,
  drop_na = TRUE
)

Value

A tibble. When format = "long": columns id, any extra identifier columns, x (numeric price), and y (numeric consumption). When format = "wide": one row per subject with prices as column names.

Arguments

data

A data frame or tibble to reshape.

format

Character. Direction of reshaping: "long" (wide to long) or "wide" (long to wide). Default "long".

id_var

Character. Name of the subject/series identifier column. Default "id".

x_var

Character. Name of the price column in long-format data. Used when format = "wide". Default "x".

y_var

Character. Name of the consumption column in long-format data. Used when format = "wide". Default "y".

price_cols

Character vector of column names in wide-format data that represent prices. Used when format = "long". Default NULL, which auto-detects: columns whose names parse as numeric become price columns, and remaining columns (besides id_var) are preserved as identifiers. If no column names parse as numeric, all non-id_var columns are treated as price columns (and x_values must be supplied).

x_values

Numeric vector of actual price values corresponding to each price column in wide-format data. Used when format = "long". Must be the same length as the number of price columns. Default NULL, which parses prices from column names.

drop_na

Logical. When format = "long", drop rows where consumption (y) is NA after pivoting? Default TRUE. A warning is issued when rows are dropped.

Details

Wide to Long (format = "long")

The function determines which columns are price columns by:

1. If price_cols is provided, those columns are used directly.

2. If price_cols is NULL, column names are tested with as.numeric(). Columns whose names successfully parse as numbers (e.g., "0", "0.5", "10") are treated as price columns. Remaining non-id_var columns are preserved as extra identifiers.

3. If no column names parse as numeric, all non-id_var columns become price columns and x_values must be supplied.

Actual numeric price values come from x_values if supplied, or from parsing column names. If column names cannot be parsed and x_values is not supplied, an error is raised.

Long to Wide (format = "wide")

Pivots long data so that each unique value in x_var becomes a column, with values from y_var. All columns except x_var and y_var are used as identifiers.

Examples

# --- Wide to long ---
# Columns named as prices (auto-parsed)
wide_num <- data.frame(
  id = 1:3,
  "0" = c(10, 8, 12), "0.5" = c(9, 7, 11), "1" = c(8, 6, 9),
  check.names = FALSE
)
pivot_demand_data(wide_num, format = "long")

# Columns with non-numeric names require x_values
wide_named <- data.frame(id = 1:2, price_1 = c(10, 8), price_2 = c(5, 4))
pivot_demand_data(wide_named, format = "long",
                  x_values = c(0, 0.5))

# --- Long to wide ---
data(apt, package = "beezdemand")
wide_apt <- pivot_demand_data(apt, format = "wide")
head(wide_apt)