In pbulsink/hockeyR: Hockey Analytics Scripts from the Community

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Run the entire script below starting at ### DRYSCRPAE ### through the end (do not run the pbp_list or pbp_df portion). After that has all been run and all functions are loaded, run the ds.compile_games function directly below this.

The "pbp_list" object is a list. [[1]] will pull the pbp specific dataframe out of the list.

I've updated all the below functions to work with the HTM links (updated event types). I've also commented out

a specific line in one of the functions that fixes a bug in Manny's original code that attempts to convert all

PHX games to ARI. Basically, you'll get all the games, but for 07-13 (I think), all PHX games will show the team

as "PHX" instead of ARI.

The "games = ..." portion can be a single game or a vector. I would recommend only scraping at most 400 games at one time.

games <- c(20303:21230)
# change here to the directory you want to write the pbp files to
dir.create("./data")


for (game in games) {
  # this should be the same as the directory above
  setwd("./data")
  # you will need to change the season to reflect what season you want to scrape
  pbp_list <- ds.compile_games(
    games = game,
    season = "20132014",
    pause = 2,
    try_tolerance = 5,
    agents = ds.user_agents
  )

  pbp_df <- pbp_list[[1]]

  # this creates a new directory inside the above directory to house each game
  # pbp I do this because I create other stats based on that one pbp file but
  # if you dont want it you can delete it
  dir.create(paste0("~/HockeyStuff/xGGameBreakdowns/2014/", as.character(game)))
  setwd(paste0("~/HockeyStuff/xGGameBreakdowns/2014/", as.character(game)))
  # this writes the pbp to a text file neccesarry if you want to read it in
  # again without scraping
  write_delim(pbp_df, as.character(game), delim = "|")
}

pbulsink/hockeyR documentation built on May 15, 2019, 4:16 p.m.