In nmmarquez/censeg: Calculate Segregation Measures from Census Data

rm(list = ls())
library(censeg)
library(ggplot2)
library(dplyr)
library(stringr)
# be sure to cache geographies to make re-calculations easier
options(tigris_use_cache=TRUE)

# get all variables which relate to the ASIAN ALONE WITH ONE ASIAN CATEGORY
noout <- capture.output(cen10_vars_df <- load_variables(2010, "sf1") %>%
    filter(str_starts(concept, "ASIAN ALONE WITH ONE ASIAN CATEGORY")) %>%
    # clean up the names of the data
    mutate(race = str_split_fixed(label, "!!", 2)[,2]) %>%
    mutate(race = str_split_fixed(race, " \\(", 2)[,1]) %>%
    filter(race != "") %>%
    select(name, race))

# restructure to a format that is amenable to the get_decennial function
cen10_vec <- cen10_vars_df$name
names(cen10_vec) <- cen10_vars_df$race

# get the data at the tract level for all metro areas with identifiers
group_df <- get_decennial_metro("tract", cen10_vec, year = 2010)
# rename variable to race because that's how the calc_functions need them
setnames(group_df, "variable", "race")
# only look at locations with an associated metro area
metro_only_df <- group_df[!is.na(cbsacode), ]

# calculate Vietnamese clusters for all metro areas in the US
viet_lclust_df <- calc_lisa(
    metro_only_df, "Vietnamese", parent_geo = "cbsatitle", year = 2010)

# Locations by their signficant number of Vietnamese clusters
viet_lclust_df[,.(N = sum((Ip < .05) & hiclust)), by="cbsatitle"][order(-N),]

nmmarquez/censeg documentation built on Sept. 12, 2020, 4:13 a.m.