Texas congressional district simulation

output.md

Texas congressional district simulation.R

library(tidyverse)

# arbitrary number
district_count = 38

# population from stephanie's figure
# https://twitter.com/evergreendata/status/1450862060972216320
population = c(
  rep("White", 40),
  rep("Latino", 39),
  rep("Black", 12),
  rep("Asian", 5),
  rep("Other", 4)
)
population = rep(population, each=district_count)

# what is the majority in a given district
# if no ethnicity is >50%, return "no majority"
get_majority = function(ethnicity) {
  majority = (table(ethnicity) %>% sort(TRUE))[1]
  ifelse(majority / length(ethnicity) <= 0.5,
         "no majority",
         majority %>% as.list() %>% names())
}

simulate = function() {
  tibble(
    ethnicity = sample(population, length(population), replace=FALSE),
    district = 0:(length(ethnicity)-1) %/% district_count
  ) %>% 
    group_by(district) %>% 
    summarize(majority = get_majority(ethnicity)) %>% 
    ungroup() %>% 
    count(majority)
}

simulated_results = lapply(1:1000, function(x) simulate() %>% mutate(iteration=x))
simulated_results = simulated_results %>% bind_rows()
simulated_results = simulated_results %>% mutate(majority = factor(majority, c("White", "Latino", "Black", "Asian", "no majority")))

ggplot(simulated_results) + 
  aes(x=n, y=fct_rev(majority), fill=majority) +
  ggdist::stat_histinterval() +
  scale_fill_manual(values = c(
    "White" = "#1B9E77",
    "Latino" = "#D95F02",
    "no majority" = "gray"
  )) +
  guides(fill = "none") +
  labs(x = NULL, y=NULL,
       title = "If everyone in Texas lived in a random location and there were no gerrymandering,\nwhat would the majority racial group (>50%) be for the districts?",
       #subtitle = "(a simulation)",
       caption = "data source: @evergreendata\nby Steve Haroz") +
  theme_minimal(18)