groupwise.R

library(tidycensus)
library(tidyverse)
library(tigris)
library(sf)
options(tigris_use_cache = TRUE)

# You'll need a Census API key to get this to work
# census_api_key("YOUR KEY HERE")

vt_tract_pop <- get_decennial(
  geography = "tract",
  variables = "P1_001N",
  state = "VT",
  year = 2020,
  geometry = TRUE
)

vt_county_pop <- vt_tract_pop %>%
  mutate(county_fips = str_sub(GEOID, 1, 5)) %>%
  group_by(county_fips) %>%
  summarize(pop = sum(value, na.rm = TRUE))

leafsync::sync(
  mapview::mapview(vt_tract_pop, zcol = "value", layer.name = "Tract population"),
  mapview::mapview(vt_county_pop, zcol = "pop", layer.name = "County population")
)

# Example 1: regular group by / summarize. Crashes my computer!
# Will probably work with a projected CRS or sf::sf_use_s2(FALSE) but will be slow
tx_income_groups <- get_acs(
  geography = "tract",
  table = "B19001",
  state = "TX",
  year = 2020,
  geometry = TRUE
) %>%
  filter(variable != "B19001_001") %>%
  mutate(bracket = case_when(
    variable > "B19001_013" ~ "Above $100k",
    TRUE ~ "Below $100k"
  )) %>%
  group_by(GEOID, bracket) %>%
  summarize(n_households = sum(estimate, na.rm = TRUE))

# Example 2: do_union = FALSE
tictoc::tic()
tx_income_groups <- get_acs(
  geography = "tract",
  table = "B19001",
  state = "TX",
  year = 2020,
  geometry = TRUE
) %>%
  filter(variable != "B19001_001") %>%
  mutate(bracket = case_when(
    variable > "B19001_013" ~ "Above $100k",
    TRUE ~ "Below $100k"
  )) %>%
  group_by(GEOID, bracket) %>%
  summarize(n_households = sum(estimate, na.rm = TRUE),
            do_union = FALSE)
tictoc::toc()

# Example 3: get tracts independently
tictoc::tic()
tx_tracts <- tracts("TX", cb = TRUE, year = 2020) %>%
  select(GEOID)

tx_income_groups <- get_acs(
  geography = "tract",
  table = "B19001",
  state = "TX",
  year = 2020,
  geometry = FALSE
) %>%
  filter(variable != "B19001_001") %>%
  mutate(bracket = case_when(
    variable > "B19001_013" ~ "Above $100k",
    TRUE ~ "Below $100k"
  )) %>%
  group_by(GEOID, bracket) %>%
  summarize(n_households = sum(estimate, na.rm = TRUE)) 

tx_income_groups_sf <- tx_tracts %>%
  left_join(tx_income_groups, by = "GEOID")
tictoc::toc()