Parse NHL boxscores from http://hockey-reference.com/boxscores using R

nhl_boxscore_scraper.R

library(XML)
# parses boxscores from hockey-reference.com. 
# Note program may take several hours to run.
# change wd to folder where data should be stored
# change months and year variables to date range preferred
setwd("~/Desktop/Hockey")

months <- c("03", "04", "10", "11", "12")
days <- as.character(seq(1,31))
for(i in 1:31) if(nchar(days[i])==1) days[i] = paste0("0",days[i])
year <- "2014"

teams <- c("CAR", "FLA", "VAN", "CGY",
           "DAL", "EDM", "TBL", "LAK",
           "NJD", "NSH", "WPG", "MTL",
           "PIT", "SJS", "WSH", "STL",
           "BUF", "BOS", "NYI", "DET",
           "CBJ", "PHI", "OTT", "COL",
           "ARI", "MIN", "NYR", "CHI",
           "TOR", "ANA")

for(month in months) {
  for(day in days) {
    for(team in teams) {
      url <- paste0("http://www.hockey-reference.com/boxscores/", year, month, day, "0", team, ".html")
      team_data <- tryCatch({readHTMLTable(url)},error=function(e)NA)
      csv_name <- paste0(year,month,day)
      if(!is.na(team_data)) {
        # first team is home -> second team is away
        x <- names(team_data)
        y <- substr(x,1,3)
        z <- sapply(teams,function(x) which(y==x)[1])
        z <- z[which(!is.na(z))]
        znames <- names(z)
        for(i in 1:length(z)) {
          write.csv(team_data[[z[i]]], file=paste0(znames[i], csv_name,".csv"))
        }
      }
    }
  }
}
player_data <- carolina_data[[13]]