ndi.R

ndi <- function(state,county){
  
  vars <- c("B17001_002", "B17001_001", "B06009_002" , 
            "B06009_001", "B09008_011", "B09008_001",
            "B08124_002", "B08124_001", "B25014_005",
            "B25014_006", "B25014_007", "B25014_011", 
            "B25014_012", "B25014_013", "B25014_001", 
            "B19058_002", "B19058_001", "C23002A_008E", 
            "C23002A_021E", "C23002A_006E", "C23002A_019E", 
            "B19001_002","B19001_003", "B19001_004",
            "B19001_005", "B19001_006", "B19001_001")
  
  if (missing(county)) {
    county <- NULL
  } else {
    county <- county
  }
  
  acs_data <- tidycensus::get_acs(geography = "tract", 
                                  variables = vars,
                                  state = state,  
                                  county = county,
                                  output = "wide") %>%
    dplyr::mutate(pct_poverty = B17001_002E / B17001_001E,
           pct_noHS = B06009_002E / B06009_001E,
           pct_FHH = B09008_011E / B09008_001E,
           pct_mgmt = B08124_002E /  B08124_001E, 
           pct_crowd =  (B25014_005E + B25014_006E + B25014_007E + 
                           B25014_011E + B25014_012E + B25014_013E) / B25014_001E,
           pct_pubassist = B19058_002E / B19058_001E,
           pct_unempl = (C23002A_008E + C23002A_021E) / (C23002A_006E + C23002A_019E),
           pct_under30K =( B19001_002E + B19001_003E + B19001_004E +
                             B19001_005E +  B19001_006E) / B19001_001E)

  values  <-  acs_data %>% 
    dplyr::select(pct_poverty,
                  pct_noHS,
                  pct_FHH,
                  pct_mgmt,
                  pct_crowd,                              
                  pct_pubassist, 
                  pct_unempl,
                  pct_under30K) %>% 
    as.matrix()

  values[is.nan(values)] <- 0
  
  ND <- psych::principal(values,
                         nfactors = 1)          
  
  NDI <- cbind(acs_data,
               ND$scores) 
  
  NDI <- NDI %>% 
    dplyr::select(NAME, 
                  GEOID, 
                  PC1, 
                  57:64) %>% 
    tidyr::separate(NAME, 
                    into = c("Tract", "County", "State"), 
                    sep = ",") %>% 
    dplyr::mutate(County = stringr::str_trim(County), 
                  State = stringr::str_trim(State)) %>%
    rename(NDI = PC1)  %>% 
    dplyr::as_tibble() 
  
  return(NDI)
  
}