process-games: Given downloaded NHL games, produce event tables with players...
In nhlscrapr: Compiling the NHL Real Time Scoring System Database for easy use in R

Description Usage Arguments Details Value Author(s)

Produces game tables from the NHL source files.

  process.single.game (season="20122013",
                       gcode="20001",
                       rdata.folder="nhlr-data",
                       override.download=FALSE,
                       save.to.file=TRUE,
                       ...)
  process.games       (games=full.game.database(),
                       rdata.folder="nhlr-data",
                       override.download=FALSE)
  retrieve.game       (season="20122013",
                       gcode="20001",
                       rdata.folder="nhlr-data",
                       force=TRUE,
                       ...)
  compile.all.games   (rdata.folder="nhlr-data",
                       output.folder="source-data",
                       new.game.table=NULL,
                       seasons=NULL,
                       verbose=FALSE,
                       override.days.back=NULL,
                       date.check=FALSE,
                       ...)

`season`	A character string for the two years specifying an NHL season.
`gcode`	The five-digit ID number for a particular NHL game.
`rdata.folder`	The location within the current directory to which to save the downloaded files. Will be created if it does not exist.
`output.folder`	The location within the current directory where compiled files will be saved. Will be created if it does not exist.
`games, new.game.table`	A game database, such as the one produced by full.game.database().
`override.download`	Re-download the game files whether or not they are currently locally available.
`save.to.file`	Save the game object to file.
`force`	If TRUE, reprocesses the single game file.
`seasons`	If selected, defaults to a game table with only these seasons, of the form "20142015".
`date.check`	Try and download files without known dates.
`override.days.back`	Force download for those days previous to this one.
`verbose`	Print extra output information.
`...`	Arguments to pass to download.single.game().

This group of functions takes the downloaded HTML, image and JSON files and produces an event table for each game, particularly focusing on the players on the ice during each event.

process.single.game() produces an object called game.info that contains several pieces of information, primarily the data.frame playbyplay, with each column representing an event. Columns are as follows: event: The event number as recorded in the game. period, seconds: The period of the game and the elapsed time in seconds. a1,...,a6: The numbers and names of the away team players on the ice. h1,...,h6: The numbers and names of the home team players on the ice. ev.team: The team that registered the event in question. ev.player.1, ev.player.2, ev.player.3: The players involved in the event. distance: The distance in feet from the relevant goal for a shot on goal, missed shot or goal. type: Further information on the event, either shot type or penalty type. homezone: The zone in which the event took place, from the perspective of the home team. xcoord, ycoord: If available, the (x,y) location of a shot on goal or hit.

process.games() runs this routine and saves all game files to disk.

retrieve.game() retrieves the processed file from disk if it exists, and if not, it runs process.single.game() first.

compile.all.games() does everything that's needed to construct the entire record. This will be fastest if all games have already been downloaded and processed, possibly in parallel.

A.C. Thomas <act@acthomas.ca>

nhlscrapr documentation built on May 29, 2017, 12:41 p.m.