electionresults.R

library(rvest)
library(dplyr)
library(tidyr)
library(ggplot2)
library(stringr)
library(ggbeeswarm)
html = read_html("https://www.electionresults.govt.nz/electionresults_2017/electorate-status.html")
lnks <- (html %>% html_nodes("a") %>% html_attr("href"))[25:95]
urls <- tibble(lnks) %>% separate(lnks, into=c("before","url"), sep=2)
x <- urls$url[1]
candy <- function(x){
  electlink <- paste0("https://www.electionresults.govt.nz/electionresults_2017/", x)
  election <- read_html(electlink)
  eresult <- html_table(html_nodes(election, "table"), fill=TRUE)
  candidates <- eresult[[3]]
  candidates$electorate <- x
  return(candidates)
  }
eresult <- bind_rows(lapply(urls$url, candy))
eresult$votes <- as.numeric(gsub("[^1234567890]","",eresult$Candidates))
eresult$enum <- as.numeric(gsub("[^1234567890]","",eresult$electorate))
eresult$letter1 <- sapply(eresult$Candidates,
                          function(x){unlist(strsplit(x,
                                                      split = ""))[1]})
eresult %>% filter(!is.na(votes), enum < 65,
                   !str_detect(Candidates, fixed("Candidate ")),
                   !str_detect(Candidates, fixed("TOTAL:"))) %>% 
  group_by(electorate) %>%
  mutate(percent = 100 * votes/sum(votes)) %>% ungroup() %>%
  mutate(letter1 = factor(letter1, levels=LETTERS)) %>%
    ggplot(aes(x=as.numeric(letter1), y=percent)) + 
  geom_quasirandom(alpha=0.3) + geom_smooth() + 
  ggtitle("General electorate candidate vote share 2017 NZ general election,
by first letter of surname") + xlab("First letter") +
  scale_x_continuous(breaks=1:26, labels=LETTERS) +
  theme_minimal()