data-raw/read_non-census_data.R

In bwilden/bper: Bayesian Prediction for Ethnicity and Race

# First names ethnorace data
# Tzioumis (2018); https://www.nature.com/articles/sdata201825
first_names <- readxl::read_xlsx(here::here("data-raw", "firstnames.xlsx"),
                                 sheet = "Data") %>%
  rename(pctaapi = pctapi)

# American National Election Survey for political party ethnorace data
# https://electionstudies.org/data-center/anes-time-series-cumulative-data-file/
anes <- haven::read_dta(here::here("data-raw", "anes_timeseries_cdf.dta")) %>%
  filter(VCF0004 >= 1996) %>%
  select(year = VCF0004, ethnorace = VCF0105a, party = VCF0303) %>%
  mutate(party = if_else(is.na(party), 2, as.numeric(party)),
         party = case_when(party == 1 ~ "DEM",
                          party == 2 ~ "UNA",
                          party == 3 ~ "REP"),
         ethnorace = case_when(ethnorace == 1 ~ "white",
                               ethnorace == 2 ~ "black",
                               ethnorace == 3 ~ "aapi",
                               ethnorace == 4 ~ "aian",
                               ethnorace == 5 ~ "hispanic",
                               ethnorace == 6 ~ "other",
                               TRUE ~ NA_character_),
         year_group = case_when(year <= 2005 ~ "2000",
                                year <= 2015 ~ "2010",
                                TRUE ~ "2020")) %>%
  group_by(year_group, party) %>%
  summarise(white = sum(ethnorace == "white", na.rm = T),
            black = sum(ethnorace == "black", na.rm = T),
            aapi = sum(ethnorace == "aapi", na.rm = T),
            aian = sum(ethnorace == "aian", na.rm = T),
            hispanic = sum(ethnorace == "hispanic", na.rm = T),
            other = sum(ethnorace == "other", na.rm = T)) %>%
  ungroup()

# State code concordance file
state_codes <- readr::read_csv(here::here("data-raw", "state_code_conc.csv")) %>%
  mutate(state_code = str_pad(STATE, width = 2, pad = "0")) %>%
  select(state_code, state = STUSAB)

# Possible ethnorace outputs
ethnorace_set <- c("aian", "aapi", "black", "hispanic", "other", "white")

# Possible input data columns
bper_vars <- c("last_name", "first_name", "sex_age", "age", "sex", "party", "multi_unit",
               "county", "state", "place", "zip", "tract", "block", "district")

# Possible geographic levels
bper_geos <- c("state", "county", "zip", "place", "tract", "block", "district")


usethis::use_data(first_names, anes, state_codes,
                  bper_vars, ethnorace_set, bper_geos,
                  internal = TRUE,
                  overwrite = TRUE)