Load the South Korean Election Campaign Booklet corpus — load_campaign

Loads the election campaign booklet dataset (2000–2022) as a data.table. The dataset contains manifesto booklets from 49,678 individual candidates across presidential, National Assembly, and local elections.

Usage

load_campaign_booklet(path = NULL, cache = TRUE, refresh = FALSE)

Arguments

path: Character; path to a local CSV file. If NULL (default), the bundled dataset is used.
cache: Logical; if TRUE (default), the data is cached as an RDS file in the user's cache directory for faster subsequent loads.
refresh: Logical; if TRUE, any existing cache is ignored and the source CSV is re-read. Defaults to FALSE.

Value

A data.table::data.table with the following columns:

Column	Description
`date`	Election date
`name`	Candidate name
`region`	Region (province/metropolitan city)
`district`	Electoral district
`office_id`	Office type identifier
`office`	Office type (e.g., `national_assembly`)
`party`	Party name (Korean)
`party_eng`	Party name (English)
`result`	Election result
`sex`	Sex of the candidate
`age`	Age at the time of election
`text`	Full text of the campaign booklet
`filtered`	Filtered/preprocessed text indicator

See data_dictionary.md for the complete column reference, and the Data Descriptor (Tables 4–8) for detailed variable mappings.

Details

For full methodology and variable descriptions, see the Data Descriptor: Lim, T.H. (2025). Scientific Data, 12, 1030. doi:10.1038/s41597-025-05220-4

Examples

if (FALSE) { # \dontrun{
# Load with default caching
cb <- load_campaign_booklet()

# Load from a specific file
cb <- load_campaign_booklet(path = "path/to/my_data.csv")

# Force refresh the cache
cb <- load_campaign_booklet(refresh = TRUE)
} # }