Skip to contents

Download population data (processed asset or primary source)

Source: R/download_pop_data.R
download_pop_data.Rd

Fetches population denominators that are too large to bundle, for either scheme. raw = FALSE (the default) fetches the narcan-processed Release-asset parquet(s) (state and/or county) for the scheme and verifies each sha256 against the shipped manifest. raw = TRUE instead fetches the PRIMARY published source file(s) for the scheme – Census PEP for "single", SEER for "bridged" – verbatim.

Usage

download_pop_data(
  scheme = c("single", "bridged"),
  raw = FALSE,
  refresh = FALSE,
  dest = NULL
)

Arguments

scheme

denominator scheme: "single" (default) or "bridged"

raw

if FALSE (default), fetch the narcan-processed Release-asset parquet(s) for the scheme; if TRUE, fetch the primary published source file(s) only (not a full from-scratch reproduction of the backfill – see Details)

refresh

re-download even if a cached copy exists

dest

optional destination directory (default: the narcan cache)

Value

the local path(s) of the fetched file(s), invisibly

Details

raw = TRUE returns the recent-vintage source only and does NOT fully reproduce the multi-vintage backfill: the single-race processed data (2000-2024) pools 2000-2019 Census intercensal inputs and the bridged data (1969-2024) draws on more than one SEER extract, none of which the manifest lists. For a complete from-scratch rebuild use the data-raw/ builders. raw = TRUE emits a one-time message saying as much, so an incomplete pull never silently looks complete.

Files cache under tools::R_user_dir("narcan", "cache"). Only bulk flat-files are used (never the Census Data API); requests carry a generic narcan/<version> User-Agent and no personal identifiers.

Examples

if (FALSE) { # \dontrun{
# processed parquet(s) for the scheme (analysis-ready)
download_pop_data(scheme = "single")
# primary published source file(s) (recent vintage only; see Details)
download_pop_data(scheme = "single", raw = TRUE)
} # }