Skip to contents

download ACS 5year data from Census API, at block group resolution (slowly if for entire US)

Usage

acs_bybg(
  variables = c(pop = "B01001_001"),
  table = NULL,
  cache_table = FALSE,
  year = NULL,
  output = "wide",
  state = stateinfo$ST,
  county = NULL,
  zcta = NULL,
  geometry = FALSE,
  keep_geo_vars = FALSE,
  summary_var = NULL,
  key = NULL,
  moe_level = 90,
  survey = "acs5",
  show_call = FALSE,
  geography = "block group",
  dropname = TRUE,
  ...
)

Arguments

variables

Vector of variables - see get_acs from tidycensus package

table

see get_acs from tidycensus package.

EJSCREEN-relevant key tables at block group resolution include these: acstabs <- c("B01001", "B03002", "B15002", "C16002", "C17002", "B25034", "B23025") and at tract resolution: "B18101"

cache_table

see get_acs from tidycensus package

year

e.g., 2022 see get_acs from tidycensus package, and the helper function in the EJAM package called acsendyear()

output

see get_acs from tidycensus package

state

Default is 2-character abbreviations, vector of all US States, DC, and PR.

county

see get_acs from tidycensus package

zcta

see get_acs from tidycensus package

geometry

see get_acs from tidycensus package

keep_geo_vars

see get_acs from tidycensus package

summary_var

see get_acs from tidycensus package

key

see get_acs from tidycensus package

moe_level

see get_acs from tidycensus package

survey

see get_acs from tidycensus package

show_call

see get_acs from tidycensus package

geography

"block group"

dropname

whether to drop the column called NAME

...

see get_acs from tidycensus package

Value

data.table (not tibble, and not just a data.frame)

Details

Probably requires getting and specifying an API key for Census Bureau ! (at least if query is large). see tidycensus package help

NOTES ON KEY TABLES IN ACS THAT ARE RELEVANT TO EJSCREEN:

x <- tidycensus::load_variables(2022, "acs5")

acstabs <- c("B01001", "B03002", "B15002", "C16002", "C17002", "B25034", "B23025",
             "B18101") # disability at tract resolution only
acstabs2 <- paste0(acstabs, "_")
mytables <- data.table::rbindlist(lapply(acstabs2, function(z) {x[substr(x$name,1,7) %in% z, ][1,]}))
print(mytables)

         name            label                                                            concept   geography
       <char>           <char>                                                             <char>      <char>
  1: B01001_001 Estimate!!Total:                                                         Sex by Age block group
  2: B03002_001 Estimate!!Total:                                  Hispanic or Latino Origin by Race block group
  3: B15002_001 Estimate!!Total: Sex by Educational Attainment for the Population 25 Years and Over block group
  4: C16002_001 Estimate!!Total:    Household Language by Household Limited English Speaking Status block group
  5: C17002_001 Estimate!!Total:             Ratio of Income to Poverty Level in the Past 12 Months block group
  6: B25034_001 Estimate!!Total:                                               Year Structure Built block group
  7: B23025_001 Estimate!!Total:             Employment Status for the Population 16 Years and Over block group
  8: B18101_001 Estimate!!Total:                                    Sex by Age by Disability Status       tract

  # see details of the variables

x[substr(x$name,1,7) %in% "B01001_", ] |> print(n=50)
x[substr(x$name,1,7) %in% "B03002_", ] |> print(n=50)
x[substr(x$name,1,7) %in% "B15002_", ] |> print(n=50)
x[substr(x$name,1,7) %in% "C16002_", ] |> print(n=50)
x[substr(x$name,1,7) %in% "C17002_", ] |> print(n=50)
x[substr(x$name,1,7) %in% "B25034_", ] |> print(n=50)
x[substr(x$name,1,7) %in% "B23025_", ] |> print(n=50)
x[substr(x$name,1,7) %in% "B18101_", ] |> print(n=50)

 # disability is by tract only:

 cbind(unique(grep("disab", x$concept, value = T, ignore.case = T) ))
 # x[substr(x$name,1,6) %in% "B18101" & x$geography == "block group", ] |> print(n=50) # none
 x[substr(x$name,1,7) %in% "B18101_"  , ] |> print(n=50)

Examples

# \donttest{
## All states, full table
# newvars <- acs_bybg(table = "B01001")

## One state, some variables
newvars <- acs_bybg(c(pop = "B01001_001", y = "B01001_002"), state = "DC")

## Format new data to match rows of blockgroupstats

setnames(newvars, "GEOID", "bgfips")
dim(newvars)
newvars <- newvars[blockgroupstats[,.(bgfips, ST)], ,  on = "bgfips"]
dim(blockgroupstats)
dim(newvars)
newvars
newvars[ST == "DC", ]

## Calculate a new indicator for each block group, using ACS data

mystates = c("DC", 'RI')
newvars <- acs_bybg(variables = c("B01001_001", paste0("B01001_0", 31:39)),
  state = mystates)
setnames(newvars, "GEOID", "bgfips")
newvars[, ST := fips2state_abbrev(bgfips)]
names(newvars) <- gsub("E$", "", names(newvars))

# provide formulas for calculating new indicators from ACS raw data:
formula1 <- c(
 " pop = B01001_001",
 " age1849female = (B01001_031 + B01001_032 + B01001_033 + B01001_034 +
      B01001_035 + B01001_036 + B01001_037 + B01001_038 + B01001_039)",
 " pct1849female = ifelse(pop == 0, 0, age1849female / pop)"
 )
newvars <- calc_ejam(newvars, formulas = formula1,
  keep.old = c("bgid", "ST", "pop", 'bgfips'))

newvars[, pct1849female := round(100 * pct1849female, 1)]
mapfast(newvars[1:10,], column_names = colnames(newvars),
     labels = gsub('pct1849female', 'Women 18-49 as % of residents',
              gsub('age1849female', 'Count of women ages 18-49',
             fixcolnames(colnames(newvars), 'r', 'long'))))

# }