R/scrape.population.distribution.R

In mRchmadness: Numerical Tools for Filling Out an NCAA Basketball Tournament Bracket

#' Scrape the average rate of teams being picked to win across all ESPN brackets
#'
#' @param year the numeric year to scrape, either 2016 or 2017
#' @param sex either 'mens' or 'womens'
#' @return data.frame giving percentage of population picking each team in each round
#' @examples
#' populationDistribution = scrape.population.distribution(2017)
#' @export
#' @author eshayer
scrape.population.distribution = function(year, sex = c('mens', 'womens')) {
  sex = match.arg(sex)
  `%>%` = dplyr::`%>%`

  if (!(year %in% c(2016, 2017))) {
    stop(paste0('The year ', year, ' is not available'))
  }

  url = paste0('http://games.espn.com/tournament-challenge-bracket',
               ifelse(sex == 'mens', '', '-women'), '/', year,
               '/en/whopickedwhom')
  
  cells = xml2::read_html(url) %>%
    rvest::html_nodes('table.wpw-table td')
  
  names = cells %>%
    rvest::html_nodes('span.teamName') %>%
    rvest::html_text(trim = TRUE)
  
  probabilities = cells %>%
    rvest::html_nodes('span.percentage') %>%
    rvest::html_text(trim = TRUE) %>%
    substr(0, nchar(.) - 1) %>%
    as.numeric / 100
  
  round = rep(1:6, 64)
  
  data.frame(names = names, probabilities = probabilities, round = round)
}