census: commute mode by state leg dist

commute_mode.R

library(tidycensus)
library(tidyverse)

# find the variables we care about
acs5_2019 <- tidycensus::load_variables(year = 2019, dataset = 'acs5', cache = T)

# ok here they are
my_vars <- paste0("B08006_", sprintf('%03d', 1:17))
# select just the most important top level responses
# (e.g. no public transportation break out)
my_vars_subset <- my_vars[c(1,3,4,8,14:17)]
my_var_details <- 
  acs5_2019 %>% 
  filter(name %in% my_vars_subset)

# pull data from census using my stored API key
result_dat <-
  tidycensus::get_acs(geography = "state legislative district (lower chamber)",
                      year = 2019,
                      variables = my_vars_subset,
                      state = 'NY',
                      survey = 'acs5')

# subset to just dist 38
dist38 <- result_dat %>% filter(grepl('District 38', NAME))

# add nice category details
dist38 <- merge(dist38,
                select(my_var_details, c(name, label)),
                by.x='variable', by.y='name')

# calc proportion of total and clean up labels
dist38 <- dist38 %>%
  mutate(pct = estimate / dist38$estimate[1],
         clean_name = gsub('^.*(!!){1,2}', '', label))
dist38$clean_name <- gsub(':', '', dist38$clean_name)

# chart
dist38[-1, ] %>%
  arrange(pct) %>%
  ggplot(aes(x=fct_inorder(clean_name), y=pct)) +
  geom_col(fill='red') +
  geom_text(aes(label=scales::percent(pct, accuracy = 1)), hjust = -0.5) +
  theme_light() +
  scale_y_continuous(labels = scales::percent, limits = c(0,1)) +
  coord_flip() +
  labs(title='Journey To Work, NY State Leg Distr 38',
      subtitle = 'ACS 5 year 2019',
      x='mode', y='Percent')