Note as of May 15th 2016 - I'm in the process of updating the package to submit to CRAN. I have made a number of changes to code and data formats. If you find something doesn't work anymore, please submit an issue.
This R package is mainly a repository for complete soccer datasets, along with some built-in functions for analyzing parts of the data. Currently I include three English ones (League data, FA Cup data, Playoff data - described below) and some European leagues (Spain, Germany, Italy, Holland).
Free to use for non-commerical use. Compiled by James Curley.
Please cite as: James P. Curley (2016). engsoccerdata: English Soccer Data 1871-2015. R package version 0.1.4
If you do use it on any publications, blogs, websites, etc. please note the source (i.e. me!). Also, if you do use it - I would love to see any analysis produced from it etc. Of course, I accept no responsibility for any error that may be contained herewithin.
Contact details: jc3181 AT columbia DOT edu
To install this directly into R.
library(devtools)
install_github('jalapic/engsoccerdata', username = "jalapic")
library(engsoccerdata)
data(package="engsoccerdata") # lists datasets currently available
If you get an error message like this one
Error in curl::curl_fetch_memory(url, handle = handle) :
Problem with the SSL CA cert (path? access rights?)
which has happened on occasions for me, try this:
library(RCurl)
library(httr)
set_config( config( ssl_verifypeer = 0L ) )
library(devtools)
install_github('jalapic/engsoccerdata', username = "jalapic")
library(engsoccerdata)
Last update: 15 May 2016, v0.1.5
I am about to submit this package to CRAN. I would love help in collating more results. If anyone wants to work on a particular league or competition please let me know. These are the things I'd like to work on:
data-raw
that has league cup data up to 2013/14 but it needs error checkingdata-raw
but it needs checking also.maketable
family of funcitonsSome built-in functions:
games_between.r - returns all games ever played between two teams
games_between_sum.r - returns the summary of results between any two teams
alltimerecord.r - returns the all time record of any team in the league
score_most.r - returns the team who has been involved in the most games of each scoreline
score_teamX.r - Lists all matches that a team has played in that ended in a scoreline
score_team.r - List all occurrences of a specific scoreline for a specific team
scoreline_by_team.r - How often each team has a won,lost,drawn by a scoreline?
totalgoals_by_team.r - Return all instances of a team being involved in a game with n goals
ngoals.r - Return number of times a team has scored n goals
n_offs.r - Will return the scorelines that have occurred n number of times
opponentfreq.r - Return how often a team has played each opponent
opponents.r - number of unique opponents for all teams in total or by tier
bestwins.r - best wins for each team
worstlosses.r - worst losses for each team
maketable.r - make a league table - probably the quickest way to make a league table
maketable_eng.r - make a league table that follows the tie-breaking and points procedures for each season.
all top 4 tier games ever played 1888-2016
PL = Premier League
1888/9-1891/2 FL Division 1
In the csv file, I've used divisions 1,2,3,3a,3b, 4 as the notation I've also used tier 1,2,3,4 - to refer to 3,3a & 3b all belonging to tier 3
Dataset includes:
teams that dropped out half way through a season: - 1919 Leeds City - 1931 Wigan Borough - 1961 Accrington Stanley
includes 1919 Port Vale who replaced Leeds City mid-season
The truncated 1939/40 season is in a separate file england1939.csv
Team Names used in the file are those that are currently used: e.g. Small Heath are Birmingham City, Ardwick are Manchester City, etc.
The modern Accrington Stanley are 'Accrington' to distinguish from original Accrington Stanley and earlier Accrington FC
This was a pain to put together. It contains every single FA Cup tie (whether played or not) from the first inception of the competition in 1871 to the 2015/16 season. It does not contain pre-qualifying rounds (yet). It is best to describe each variable name in turn to give more information:
Important notes to above:
I have tried to make the dataset as complete as possible. The FA Cup data is difficult as some of it is just unobtainable. For instance, I have added venues and attendances for all semis and finals and have included this information sporadically wherelse I was able to get it. I have not done a systematic application of this to early rounds. Several games in the FA Cup are played at neutral grounds or even the visiting team is allowed to play at home (e.g. if a minnow plays a big team). I have not managed to systematically check this. Also, there was a trend to play 2nd and 3rd and 4th replays at neutral venues. This could be systematically checked but I have not yet. Further, I think I have all games that ever ended in penalties added in correctly.
Finally, team names. There are great disputes about which teams branch off from which teams in history and who should have shared history. I have tried to be consistent in naming teams with their most current name throughout (e.g. Millwall Rovers, Millwall Athletic and Millwall are all listed as the current name - Millwall), or the name that they used when they stopped playing (e.g. Mitchell St. George's are always listed as Birmingham St. George's). I have also tried to follow the same team name format as in england.csv - I think the three Accrington teams may be the only one I need to re-edit for this purpose.
Please refer to the spainliga rpubs below for further information.
I've just added complete all top tier results for Holland (1956-2016), Germany (1963-2016) and Italy (1934-2016). These dataframes contain all league results played in regular season. They don't yet include relegation/promotion playoff fixtures. Further, I have not yet completed all final checks of the data. I believe they are error free - but if others want to test and check, I'd welcome this.
Any help in improving the quality of these datasets is appreciated.
(note as of May 2015, the code in these may need to change to reflect the change in names of datasets and some functions) - http://rpubs.com/jalapic/daygoals #goal scoring trends on unqiue dates in soccer history - http://rpubs.com/jalapic/facuplast8 #quick walkthrough of some of the FA Cup data - http://rpubs.com/jalapic/gpg #very quick look at id-ing breakpoints in English scoring trends - http://rpubs.com/jalapic/gamebygame #plotting game by game trends across seasons - http://rpubs.com/jalapic/seasons #visualizing season to season changes in top tier performance - http://rpubs.com/jalapic/laliga #visualizing historical Spanish La Liga data
Oliver Roeder and I have written several articles for fivethirtyeight using these data:
Also this piece on league inequality:
(listing them here so I don't forget them)
Dec 2014 - Profile of this dataset and me in "FourFourTwo" Magazine - https://www.scribd.com/doc/246229712/FourFourTwo-UK-2014-12
Mar 12th 2015 - Some research on strange results occuring in a row discussed on Guardian's Football Weekly https://soundcloud.com/guardianfootballweekly/football-weekly-extra-chelseas-champion-league-campaign-goes-down-the-tube
Jul 30th 2015 - Piece by Sky Sports on homefield advantage and these data - http://www.skysports.com/football/news/11661/9829828/home-advantage-is-not-as-important-as-it-once-was-finds-sky-sports-study
Nov 28th 2015 - I discuss home-field advantage on NPR's "Only a Game" - http://onlyagame.wbur.org/2015/11/28/home-field-advantage-epl-curley
May 4th 2016 - I discussed this dataset and Leicester City on BBC Radio 5's "Up All Night" - unfortunately no audio - I got cut short because Ted Cruz decided to quit the Republican nomination.
May 17th 2016 - Piece by John Murdoch at the Financial Times on Leicester City's unique season - https://ig.ft.com/sites/leicester-premier-league-champions/
More in depth analysis by Simon on David Sumpter's Collective Behavior blog: - http://www.collective-behavior.com/liverpool-is-still-the-most-successful-english-club-team-but-for-how-long/ - http://www.collective-behavior.com/how-the-big-four-made-football-predictable/
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.