Convert data frame from long tidy format to wider format

The resulting data frame is a simpler form of the document feature matrix used by other packages. my_df |> expand_documents(percent = FALSE, sort = FALSE) compares to my_df |> count(doc_id, word) |> tidytext::cast_dfm(doc_id, word, n), but it is not equivalent. The latter prepares a DFM to be used with the quanteda package.

Usage

expand_documents(
  data,
  feature = word,
  by = doc_id,
  percent = TRUE,
  sort = TRUE,
  columns = NULL
)

Arguments

data

A tidy data frame, potentially containing a column called "word"

feature

A column of words containing one word per row, to be counted for frequency

by

A column containing document grouping

percent

Whether frequencies should be converted to percentages on a per-document basis

sort

Whether to sort features by frequency

columns

The features to keep

Value

A data frame with one row per document and as many features as words.

Examples

if (FALSE) { # \dontrun{
dubliners <- get_gutenberg_corpus(2814) |>
  load_texts() |>
  identify_by(part) |>
  standardize_titles()

dubliners |>
  expand_documents()
} # }