get-election-data.R

library(plyr)
library(XML)
library(uuid)
library(reshape2)

results <- ldply(states, function(state) {
  url <- "http://www.electionprojection.com/latest-polls/%s-presidential-polls-trump-vs-clinton-vs-johnson-vs-stein.php"
  state.fmt <- gsub(" ", "-", tolower(state))

  url.state <- sprintf(url, state.fmt)
  print(url.state)
  r <- readHTMLTable(url.state, stringsAsFactors=FALSE)[[3]]
  r$state <- state
  r$id <- 1:nrow(r)
  cols <- c("Dates", "Firm", "state", "Clinton", "Trump", "Johnson", "id")
  r <- r[2:nrow(r),][,cols]
  r <- melt(r, id=c("Dates", "Firm", "state", "id"), variable.name="candidate", value.name="vote")

  names(r) <- c("date", "poll", "state", "id", "candidate", "vote")
  r$race <- ""
  cols <- c("date", "race", "state", "poll", "candidate", "vote", "id")
  r <- r[,cols]
  r$vote <- as.numeric(r$vote)
  r
})