1. Introduction

We’ll walk through the analysis of the Bundesliga 22-23 football match results. We’ll load the data, preprocess it, calculate Elo ratings, and generate the points for each team in a final standings format.

2. Load necessary libraries

We start by loading the required libraries, including tidyverse for data manipulation and worldfootballR for accessing football match data.

library(tidyverse)
## Warning: package 'tidyverse' was built under R version 4.2.3
## Warning: package 'ggplot2' was built under R version 4.2.3
## Warning: package 'tibble' was built under R version 4.2.3
## Warning: package 'tidyr' was built under R version 4.2.3
## Warning: package 'purrr' was built under R version 4.2.3
## Warning: package 'dplyr' was built under R version 4.2.3
## Warning: package 'stringr' was built under R version 4.2.3
## Warning: package 'lubridate' was built under R version 4.2.3
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.2     ✔ readr     2.1.4
## ✔ forcats   1.0.0     ✔ stringr   1.5.0
## ✔ ggplot2   3.4.3     ✔ tibble    3.2.1
## ✔ lubridate 1.9.2     ✔ tidyr     1.3.0
## ✔ purrr     1.0.2     
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(worldfootballR)
## Warning: package 'worldfootballR' was built under R version 4.2.3

3. Load Bundesliga 2023 data

We load the Bundesliga 2023 data, specifying the country, gender, tier, and season-end year.

Bundesliga2023 <- worldfootballR::fb_match_results(
  country = 'GER',
  gender = 'M',
  tier = "1st",
  season_end_year = 2023
)

3.1 View the structure of the data

We examine the structure of the loaded data.

str(Bundesliga2023)
## 'data.frame':    308 obs. of  20 variables:
##  $ Competition_Name: chr  "Fußball-Bundesliga" "Fußball-Bundesliga" "Fußball-Bundesliga" "Fußball-Bundesliga" ...
##  $ Gender          : chr  "M" "M" "M" "M" ...
##  $ Country         : chr  "GER" "GER" "GER" "GER" ...
##  $ Season_End_Year : int  2023 2023 2023 2023 2023 2023 2023 2023 2023 2023 ...
##  $ Round           : chr  "Regular season" "Regular season" "Regular season" "Regular season" ...
##  $ Wk              : chr  "1" "1" "1" "1" ...
##  $ Day             : chr  "Fri" "Sat" "Sat" "Sat" ...
##  $ Date            : Date, format: "2022-08-05" "2022-08-06" ...
##  $ Time            : chr  "20:30" "15:30" "15:30" "15:30" ...
##  $ Home            : chr  "Eint Frankfurt" "Wolfsburg" "Augsburg" "Union Berlin" ...
##  $ HomeGoals       : num  1 2 0 3 3 1 1 1 3 0 ...
##  $ Home_xG         : num  0.9 1 0.7 1 1.9 0.9 1.8 0.8 4.1 0.7 ...
##  $ Away            : chr  "Bayern Munich" "Werder Bremen" "Freiburg" "Hertha BSC" ...
##  $ AwayGoals       : num  6 2 4 1 1 2 0 1 1 3 ...
##  $ Away_xG         : num  4 0.8 1.3 0.3 0.1 1.4 1.3 1.6 0.2 3.2 ...
##  $ Attendance      : num  51500 28015 25310 22012 49659 ...
##  $ Venue           : chr  "Deutsche Bank Park" "Volkswagen Arena" "WWK Arena" "Stadion An der Alten Försterei" ...
##  $ Referee         : chr  "Deniz Aytekin" "Sascha Stegemann" "Bastian Dankert" "Marco Fritz" ...
##  $ Notes           : chr  "" "" "" "" ...
##  $ MatchURL        : chr  "https://fbref.com/en/matches/6a1c18c4/Eintracht-Frankfurt-Bayern-Munich-August-5-2022-Bundesliga" "https://fbref.com/en/matches/1591be13/Wolfsburg-Werder-Bremen-August-6-2022-Bundesliga" "https://fbref.com/en/matches/3c5c48d7/Augsburg-Freiburg-August-6-2022-Bundesliga" "https://fbref.com/en/matches/58e704e1/Union-Berlin-Hertha-BSC-August-6-2022-Bundesliga" ...

3.2 Change the ‘Wk’ column to numeric

We convert the ‘Wk’ column to a numeric data type.

Bundesliga2023$Wk <- as.numeric(Bundesliga2023$Wk)

3.3 Rename columns for better readability

We rename columns to make them more understandable and descriptive.

Bundesliga2023 <- Bundesliga2023 %>%
  rename(
    team = Home,
    opp = Away,
    teamScore = HomeGoals,
    oppScore = AwayGoals,
    round = Wk,
    Match_ID = Round
  )

3.4 Remove Play off fixtures

We only want to include the German Regular Season, not the DFB Pokal rounds

Bundesliga2023 <- Bundesliga2023 %>%
  filter(Match_ID == "Regular season")

3.5 Select columns of interest

Tidying up the data frame

Bundesliga2023 <- Bundesliga2023 %>% 
  select('round','team','opp', 'teamScore', 'oppScore')

4. New Column that provides a 1= win, 0.5 = Draw and 0 = Loss

Setting up the foundation for the Elo Model

Bundesliga2023 <- Bundesliga2023 %>% 
  mutate(score = ifelse(teamScore > oppScore, 1, 
                        ifelse(teamScore < oppScore, 0, 0.5)))

5. Data Transformation

bn_new <- Bundesliga2023 %>%
  mutate(normalized = purrr::map2_chr(team, opp, ~paste(sort(c(.x, .y)), collapse = ""))) %>%
  group_by(round, normalized) %>%
  summarise(team = dplyr::first(team),
            opp = dplyr::first(opp)) %>%
  select(-normalized)
## `summarise()` has grouped output by 'round'. You can override using the
## `.groups` argument.

6. Merge with the original data

bn_new <- merge(bn_new, Bundesliga2023)

7. Arrange the data by round

bn_new <- bn_new %>%
  arrange(round)

8. Final Product

This produces our Elo Ratings for each team.

BN1_elo <- PlayerRatings::elo(
  bn_new[c('round', 'team','opp','score')],
  init = 2200,
  k = 27,
  history = T)

9. Additional Presentation of Data (If you choose to)

Create a new Data frame with all the values included from the Elo Model in the previous section

BN1_elo <- list(
  "Dortmund" = list(Rating = 2324, Games = 34, Win = 22, Draw = 5, Loss = 7, Lag = 0),
  "Bayern Munich" = list(Rating = 2319, Games = 34, Win = 21, Draw = 8, Loss = 5, Lag = 0),
  "RB Leipzig" = list(Rating = 2309, Games = 34, Win = 20, Draw = 6, Loss = 8, Lag = 0),
  "Union Berlin" = list(Rating = 2258, Games = 34, Win = 18, Draw = 8, Loss = 8, Lag = 0),
  "Freiburg" = list(Rating = 2245, Games = 34, Win = 17, Draw = 8, Loss = 9, Lag = 0),
  "Leverkusen" = list(Rating = 2225, Games = 34, Win = 14, Draw = 8, Loss = 12, Lag = 0),
  "Wolfsburg" = list(Rating = 2211, Games = 34, Win = 13, Draw = 10, Loss = 11, Lag = 0),
  "Eint Frankfurt" = list(Rating = 2210, Games = 34, Win = 13, Draw = 11, Loss = 10, Lag = 0),
  "Mainz 05" = list(Rating = 2191, Games = 34, Win = 12, Draw = 10, Loss = 12, Lag = 0),
  "Köln" = list(Rating = 2186, Games = 34, Win = 10, Draw = 12, Loss = 12, Lag = 0),
  "M'Gladbach" = list(Rating = 2180, Games = 34, Win = 11, Draw = 10, Loss = 13, Lag = 0),
  "Bochum" = list(Rating = 2154, Games = 34, Win = 10, Draw = 5, Loss = 19, Lag = 0),
  "Stuttgart" = list(Rating = 2150, Games = 34, Win = 7, Draw = 12, Loss = 15, Lag = 0),
  "Schalke 04" = list(Rating = 2147, Games = 34, Win = 7, Draw = 10, Loss = 17, Lag = 0),
  "Hoffenheim" = list(Rating = 2143, Games = 34, Win = 10, Draw = 6, Loss = 18, Lag = 0),
  "Augsburg" = list(Rating = 2122, Games = 34, Win = 9, Draw = 7, Loss = 18, Lag = 0),
  "Werder Bremen" = list(Rating = 2116, Games = 34, Win = 10, Draw = 6, Loss = 18, Lag = 0),
  "Hertha BSC" = list(Rating = 2109, Games = 34, Win = 7, Draw = 8, Loss = 19, Lag = 0)
)

10. Convert it to a data frame

BN1_elo_DF <- do.call(rbind, lapply(names(BN1_elo), function(team) {
  data.frame(
    Player = team,
    Rating = BN1_elo[[team]]$Rating,
    Games = BN1_elo[[team]]$Games,
    Win = BN1_elo[[team]]$Win,
    Draw = BN1_elo[[team]]$Draw,
    Loss = BN1_elo[[team]]$Loss,
    Lag = BN1_elo[[team]]$Lag
  )
}))

11. Revome all column headers

rownames(BN1_elo_DF) <- NULL

11.1 Add the New column header names in the correct order

BN1_elo_DF <- BN1_elo_DF %>% 
  select('Player' , 'Rating', 'Games', 'Win', 'Draw', 'Loss')

11.2 Replacing the column header of “Player” to “Team”

BN1_elo_DF <- BN1_elo_DF %>%
  rename(Team = Player)

12. Convert columns “Win” and “Draw” into Numeric formats

BN1_elo_DF$Win = as.numeric(BN1_elo_DF$Win)
BN1_elo_DF$Draw = as.numeric(BN1_elo_DF$Draw)

13. Calculate the points for wins and draws to display

BN1_elo_DF$Points <- 3*BN1_elo_DF$Win + 1*BN1_elo_DF$Draw

14. View the new Data frame

BN1_elo_DF
##              Team Rating Games Win Draw Loss Points
## 1        Dortmund   2324    34  22    5    7     71
## 2   Bayern Munich   2319    34  21    8    5     71
## 3      RB Leipzig   2309    34  20    6    8     66
## 4    Union Berlin   2258    34  18    8    8     62
## 5        Freiburg   2245    34  17    8    9     59
## 6      Leverkusen   2225    34  14    8   12     50
## 7       Wolfsburg   2211    34  13   10   11     49
## 8  Eint Frankfurt   2210    34  13   11   10     50
## 9        Mainz 05   2191    34  12   10   12     46
## 10           Köln   2186    34  10   12   12     42
## 11     M'Gladbach   2180    34  11   10   13     43
## 12         Bochum   2154    34  10    5   19     35
## 13      Stuttgart   2150    34   7   12   15     33
## 14     Schalke 04   2147    34   7   10   17     31
## 15     Hoffenheim   2143    34  10    6   18     36
## 16       Augsburg   2122    34   9    7   18     34
## 17  Werder Bremen   2116    34  10    6   18     36
## 18     Hertha BSC   2109    34   7    8   19     29