join_nccs_block.R

library( tigris )
library( tidyverse )
library( sf )

# set filepath of where to 
# save raw block shapefiles

path <- "xwalk/"


#####  
#####  get shapefiles from tigris 
#####  

# set boolean for whether you want to download block shapefiles
download_block <- FALSE

# get unique state fips codes (including dc)

unique_states <- fips_codes %>% 
  filter(state_code != 74) %>%
  pull(state_code) %>% 
  unique()


# only run if download_block is TRUE
if( download_block ){

# if shapefiles directory doesn't exist, create it
if( ! dir.exists(path) )
{ dir.create(path) }


#####  download block shapefiles by state
#####  abridge to only the geoid variable
#####  and save
	
walk(unique_states, function(x) {
  
  blocks(state = x, 
         year = 2010) %>% 
    select(GEOID10) %>% 
    st_write(str_c(path, "block_", x, ".geojson")) 
   })
  
}  # end if( download_block )


# list paths of all 
# saved block shapefiles

block_paths <- 
  list.files( path, full.names = TRUE )


# get state fips crosswalk

state_xwalk <- 
  fips_codes %>% 
  select( state_code, state ) %>% 
  distinct()


#####
#####  read in nccs data
#####

nccs_data <- read_csv( str_c( path, "tinybmf.csv" ) )

nccs_data_sf <- 
  st_as_sf( nccs_data, 
    coords = c( "Longitude", "Latitude" ),
    crs = 4326 ) %>% 
    select( id, 
  	    STATE, 
  	    Longitude, 
  	    Latitude ) %>%
    st_transform( "EPSG:4269" )

# if intermediate directory doesn't exist, create it
if( !dir.exists("intermediate") )
{ dir.create("intermediate") }

# if final directory doesn't exist, create it
if( !dir.exists("final") )
{ dir.create("final") }


#####
#####  joins nccs data to block data by state
#####

join_nccs <- function(my_state){
  print(my_state)

  # get state abbreviation  
  state_abbv <- state_xwalk %>% 
    filter(state_code == my_state) %>%
    pull(state)
  
  # keep only nccs data for the state we're joining to minimize memory and time
  nccs_state <- 
    filter( nccs_data_sf,
  	    STATE == state_abbv)
  
  # read in block geojson
  state_block <- 
	st_read( 
		str_c( path, 
  		       "block_",
  		       my_state,
  		       ".geojson" ))
  
  # join data and remove the geometry
  data_out <- 
	st_join( nccs_state, 
  		 state_block ) %>% 
    st_drop_geometry()
  
  # save the joined data

  filename <- str_c( "intermediate/joined_block_", my_state, ".csv" )
  write_csv( data_out, filename )
                           
}



#####
#####  iterate through all states and join data
#####

walk( unique_states, join_nccs)

# read in all joined data and append together
full_joined_dat <- 
  list.files( "intermediate", 
	      pattern = "joined_block_", 
	      full.names = TRUE ) %>% 
  map_dfr( ~read_csv( ., col_types = list(GEOID10 = col_character()) ))

# join abridged data with block info back to full dataset and write out
nccs_dat_full <- 
  left_join( nccs_data,
	    full_joined_dat, 
            by = c("id", "STATE") ) %>% 
  write_csv( "final/nccs_test.csv" )