global.R

## 01. Load libraries, set options, source functions and resources -------------
require(tidyverse)
require(tidycensus)
require(sf)
require(tigris)
require(mapview)

options(
  scipen = 999,
  digits = 4,
  tigris_class = "sf",
  tigris_use_cache = T
)

source("4_scripts/1_data-processing/01_load_spatial_resources.R")
source("3_functions/get_counties_from_cbsa.R")
source("3_functions/get_states_from_stcnty_fips.R")

## 02. Determine and input target region's FIPS/CBSA codes and places ----------

# TODO: Enter a region to look up in the line below to search for the correct CBSA FIPS 
REGION_OF_INTEREST <- "portland" 

# Use code below to determine the FIPS code for the region of interest ("CBSA" = region). 
# Look at the results and decide what the region's FIPS code is
county2msa %>% 
  filter(str_detect(cbsaname15, regex(REGION_OF_INTEREST, ignore_case = TRUE)))

# TODO: Enter the correct CBSA FIPS code found from lookup above to determine 
# the counties that are part of that CBSA
COUNTIES_TO_DOWNLOAD <- get_counties_from_cbsa(38900)

## 03. Automatically generate other variables we will need based ---------------

## Determine states from counties
STATES_TO_DOWNLOAD <- get_states_from_stcnty_fips(COUNTIES_TO_DOWNLOAD)

TRACTS.SF <- map_dfr(STATES_TO_DOWNLOAD, .f = function(STATE){
  tigris::tracts(STATE, cb = TRUE, year = 2020) %>%
    filter(substr(GEOID, 1, 5) %in% COUNTIES_TO_DOWNLOAD)})

TRACTS_HIRES.SF <- map_dfr(STATES_TO_DOWNLOAD, .f = function(STATE){
  tigris::tracts(STATE, cb = FALSE, year = 2020) %>%
    filter(substr(GEOID, 1, 5) %in% COUNTIES_TO_DOWNLOAD)})

# TARGET_EPSG <- crsuggest::suggest_top_crs(TRACTS.SF, units = "us-ft")
TARGET_EPSG <- 2913
WEB_EPSG <- 4326

## Main places in CBSA
PLACES_IN_CBSA <- tigris::places(STATES_TO_DOWNLOAD) %>%
  st_filter(., filter(tigris::counties(state = STATES_TO_DOWNLOAD, cb = T), GEOID %in% COUNTIES_TO_DOWNLOAD)) %>%
  st_transform(TARGET_EPSG) %>%
  arrange(desc(ALAND)) %>%
  head() 

## Grab GEOID/FIPS code of the primary place of interest, which is USUALLY
## the largest by land area. 
PRIMARY_PLACE <-  PLACES_IN_CBSA[1,]

TRACTS.SF <- st_transform(TRACTS.SF, TARGET_EPSG)
TRACTS_HIRES.SF <- st_transform(TRACTS_HIRES.SF, TARGET_EPSG)

## Identify a list of tracts that are within the primary place of interest
## Adjust negative distance buffer to fit desired tolerance
BUFFER_DISTANCE <- -1320

TRACTS_IN_PLACE <- TRACTS.SF %>%
  st_filter(., st_buffer(PRIMARY_PLACE, dist = BUFFER_DISTANCE)) %>%
  pull(GEOID)

## Quick view of the tracts in place
TRACTS.SF %>%
  mutate(in_primary_place = GEOID %in% TRACTS_IN_PLACE) %>%
  mapview(zcol = 'in_primary_place')

tracts_to_remove <- c('41005022208', '41051010200')

TRACTS_IN_PLACE <- setdiff(TRACTS_IN_PLACE, tracts_to_remove)

## Create new variable on whether that tract is in the primary place of interest
TRACTS.SF <- TRACTS.SF %>%
  mutate(in_primary_place = GEOID %in% TRACTS_IN_PLACE)

TRACTS_HIRES.SF <- TRACTS_HIRES.SF %>%
  mutate(in_primary_place = GEOID %in% TRACTS_IN_PLACE)