In mattle24/spatial4social: Geospatial Analysis for Social Science

Calculating segregation in cities with tidycensus, tigris, and spatial4social

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(spatial4social)
library(tidycensus)
library(tigris)
library(dplyr)
library(ggplot2)

options(tigris_class = "sf")

variable_lookup <- tibble::tibble(
  variable_name = c("white_non_hispanic", "afam_non_hispanic")
  ,variable_code = c("B03002_003", "B03002_004")
)

places_sf <- tigris::places("CA", year = 2015)
# limit to biggest cities by area to make this example run faster
places_sf <- places_sf %>%
  dplyr::top_n(5, wt = ALAND)

tracts <- tidycensus::get_acs("tract", variables = variable_lookup$variable_code,
                              year = 2015, state = "CA")

tracts_sf <- tigris::tracts("CA", year = 2015)
tracts <- tracts %>%
  dplyr::left_join(variable_lookup, by = c("variable" = "variable_code")) %>%
  dplyr::select(-NAME, -variable, -moe) %>%
  tidyr::spread(key = "variable_name", value = "estimate")
tracts_sf <- tracts_sf %>%
  dplyr::inner_join(tracts, by = "GEOID")

x <- calculate_segregation(macro_geometry_sf = places_sf, sub_geometry_sf = tracts_sf,
                          majority_pop = "white_non_hispanic", minority_pop = "afam_non_hispanic",
                          method = "dissimilarity_index")

ggplot(x) +
  geom_sf(aes(fill = dissimilarity_index)) +
  geom_sf_label(aes(label = NAME), nudge_y = 0.5, nudge_x = 0.5) +
  scale_x_continuous(limits = c(-123, -116)) +
  scale_y_continuous(limits = c(32.5, 38)) +
  coord_sf(datum = NA) +
  labs(title = "Dissimilarity Index in California Cities", x= "", y = "") +
  scale_fill_continuous(name = "Dissimilarity Index") +
  theme_minimal() +
  theme(legend.position = "bottom")