multi_geo_acs: Fetch an ACS table with multiple geography levels

In CT-Data-Haven/cwi: Prep for DataHaven's Community Wellbeing Index

Fetch an ACS table with multiple geography levels

Description

Fetch a data table from the ACS via tidycensus with your choice of geographies at multiple levels. For geographies made of aggregates, i.e. neighborhoods made of tracts or regions made of towns, the returned table will have estimates summed and margins of error calculated for the whole area. Any geographic levels that are null will be excluded.

Usage

multi_geo_acs(
  table,
  year = 2022,
  towns = "all",
  regions = NULL,
  counties = "all",
  state = "09",
  neighborhoods = NULL,
  tracts = NULL,
  blockgroups = NULL,
  pumas = NULL,
  msa = FALSE,
  us = FALSE,
  new_england = TRUE,
  nhood_name = "name",
  nhood_geoid = NULL,
  nhood_weight = "weight",
  survey = c("acs5", "acs1"),
  verbose = TRUE,
  key = NULL,
  sleep = 0,
  ...
)

Arguments

table	A string giving the ACS table number.
year	The year of the ACS table; currently defaults 2022 (most recent available).
towns	A character vector of names of towns to include; "all" (default) for all towns optionally filtered by county; or NULL to not fetch town-level table.
regions	A named list of regions with their town names (defaults NULL).
counties	A character vector of names of counties to include; "all" (default) for all counties in the state; or NULL to not fetch county-level table.
state	A string: either name or two-digit FIPS code of a US state. Required; defaults "09" (Connecticut).
neighborhoods	A data frame with columns for neighborhood name, GEOID of either tracts or block groups, and weight, e.g. share of each tract assigned to a neighborhood. If included, weighted sums and MOEs will be returned for neighborhoods. Try to match the formatting of the built-in neighborhood tables.
tracts	A character vector of 11-digit FIPS codes of tracts to include, or "all" for all tracts optionally filtered by county. Defaults NULL.
blockgroups	A character vector of 12-digit FIPS codes of block groups to include, or "all" for all block groups optionally filtered by county. Defaults NULL.
pumas	A character vector of 7-digit FIPS codes of public use microdata areas (PUMAs) to include, or "all" for all PUMAs optionally filtered by county. It's up to you to filter out any redundancies–some large towns are standalone PUMAs, as are some sparsely-population counties. Defaults NULL.
msa	Logical: whether to fetch New England states' metropolitan statistical areas. Defaults FALSE.
us	Logical: whether to fetch US-level table. Defaults FALSE.
new_england	Logical: if TRUE (the default), limits metro areas to just New England states.
nhood_name	String giving the name of the column in the data frame neighborhoods that contains neighborhood names. Previously this was a bare column name, but for consistency with changes to COG-based FIPS codes, this needs to be a string. Only relevant if a neighborhood weight table is being used. Defaults "name" to match the neighborhood lookup datasets.
nhood_geoid	String giving the name of the column in neighborhoods that contains neighborhood GEOIDs, either tracts or block groups. Only relevant if a neighborhood weight table is being used. Because of changes to FIPS codes, this no longer has a default.
nhood_weight	String giving the name of the column in neighborhoods that contains weights between neighborhood names and tract/block groups. Only relevant if a neighborhood weight table is being used. Defaults "weight" to match the neighborhood lookup datasets.
survey	A string: which ACS estimate to use. Defaults to 5-year ("acs5"), but can also be 1-year ("acs1").
verbose	Logical: whether to print summary of geographies included. Defaults TRUE.
key	String: Census API key. If NULL (default), takes the value from Sys.getenv("CENSUS_API_KEY").
sleep	Number of seconds, if any, to sleep before each API call. This might help with the Census API's tendency to crash, but for many geographies, it could add a sizable about of time. Probably don't add more than a few seconds.
...	Additional arguments to pass on to tidycensus::get_acs

Details

This function essentially calls tidycensus::get_acs() multiple times, depending on geographic levels chosen, and does minor cleaning, filtering, and aggregation. Note that the underlying tidycensus::get_acs() requires a Census API key. As is the case with other tidycensus functions, multi_geo_acs assumes this key is stored as CENSUS_API_KEY in your .Renviron. See tidycensus::census_api_key() for installation.

Value

A tibble with GEOID, name, variable code, estimate, moe, geography level, state, and year, as applicable, for the chosen ACS table.

See Also

tidycensus::census_api_key(), tidycensus::get_acs()

Examples

## Not run: 
multi_geo_acs("B01003", 2019,
  towns = "all",
  regions = list(inner_ring = c("Hamden", "East Haven", "West Haven")),
  counties = "New Haven County",
  tracts = unique(new_haven_tracts$geoid))

multi_geo_acs("B01003", 2018,
  towns = "Bridgeport",
  counties = "Fairfield County",
  neighborhoods = bridgeport_tracts
)


## End(Not run)

CT-Data-Haven/cwi documentation built on July 1, 2024, 7:45 a.m.