Dot density map of the 2018 NYC Democrat Governor Primary

nyc_primary_dot_density.R

library(rvest) # web scraping
library(tidyverse) # general data munging
library(sf) # spatial functions and stats
library(lwgeom) # not 100% sure why, but i needed to load this so `udunits` doesn't throw an error


## import data ------------

# import a geojson of assembly district outlines
nyc_ad <- st_read("http://services5.arcgis.com/GfwWNkhOj9bNBqoJ/arcgis/rest/services/nyad/FeatureServer/0/query?where=1=1&outFields=*&outSR=4326&f=geojson")

# scrape the table with assembly district level results from the Bureau of Elections
dem_gov <- read_html("https://enrweb.boenyc.us/CD22370AD0.html") %>%
  html_node(".underline") %>%
  html_table() %>%
  select("AssemDist" = X1, 
         "NixonVotes" = X3, 
         "CuomoVotes" = X5, 
         "WriteInVotes" = X7) %>%
  # clean up the table
  filter(grepl("AD", AssemDist)) %>%
  mutate(AssemDist = gsub("AD   ", "", AssemDist)) %>%
  mutate_all(as.numeric) 

# join the geometry and the results. isn't `sf` grand?
nyc_ad_results <- left_join(nyc_ad, dem_gov) %>%
  st_as_sf()

## calculating dots inside assembly districts --------------

# random rounding algo to prevent systemic bias in counting. credit: https://github.com/mountainMath/dotdensity/blob/master/R/dot-density.R
random_round <- function(x) {
  v=as.integer(x)
  r=x-v
  test=runif(length(r), 0.0, 1.0)
  add=rep(as.integer(0),length(r))
  add[r>test] <- as.integer(1)
  value=v+add
  ifelse(is.na(value) | value<0,0,value)
  return(value)
}

results_dots <- as.data.frame(nyc_ad_results) %>% 
  select(NixonVotes:WriteInVotes) %>% 
  # dividing results by 10 so that each dot represents 10 votes
  mutate_all(funs(. / 10)) %>% 
  # use the random_round algot to prevent systemic bias
  mutate_all(random_round)

# generate lat lon points inside assembly districts
# huge credit here to Paul Campbell's classic post: https://www.cultureofinsight.com/blog/2018/05/02/2018-04-08-multivariate-dot-density-maps-in-r-with-sf-ggplot2/
nyc_primary_dots <- map_df(names(results_dots), 
                  ~ st_sample(nyc_ad_results, size = results_dots[, .x], type = "random") %>%
                    st_cast("POINT") %>%
                    st_coordinates() %>%
                    as_tibble() %>%
                    setNames(c("lon","lat")) %>%
                    mutate(Candidate = .x)) %>% 
  slice(sample(1:n())) 

# set the palette
pal <- c("CuomoVotes" = "#70147A", "NixonVotes" = "#78B943", "WriteInVotes" = "#FCBB30")

# build the plot
p <- ggplot() +
  geom_sf(data = nyc_ad_results, fill = "transparent",colour = "white") +
  geom_point(data = nyc_primary_dots, aes(lon, lat, colour = Candidate), alpha = .6, size = .7) +
  scale_colour_manual(values = pal) +
  theme_void() +
  theme(panel.grid = element_line(color = "#141414"),
        plot.background = element_rect(fill = "#141414", color = NA), 
        panel.background = element_rect(fill = "#141414", color = NA),
        legend.background = element_rect(fill = "#141414", color = NA),
  )

ggsave("nyc_primary_density.png", plot = p, dpi = 400, width = 90, height = 70, units = "cm")