R/SeasonStats.R
In kosge1911/SeasonStats:
Defines functions
Seasons
allteams
Freqgoalplot
HAtable
crosstab
lplot
standingplot
team
league.table
allteams
Documented in
allteams
crosstab
Freqgoalplot
HAtable
league.table
lplot
Seasons
standingplot
team
###################################
# Code for the SeasonStats Package
###################################
#------------------------------------------------------------
# Included data sets
#------------------------------------------------------------
GermanBundesliga <- read.csv("de.1.csv", encoding = "UTF-8")
PremierLeague <- read.csv("eng.1.csv", encoding = "UTF-8")
PrimeraDivision <- read.csv("es.1.csv", encoding = "UTF-8")
#------------------------------------------------------------
# Function that shows you all teams
#------------------------------------------------------------
allteams <- function(games){
print(unique(games$Team.1))
}
#------------------------------------------------------------
# Function for standings during and at the end of the season
#------------------------------------------------------------
league.table <- function(games, ROUND = (length(unique(games$Team.2)) - 1) * 2){
# A function that creates the table of the season.
# games ... dataframe with all games of one season
# ROUND ... shows table of specific round
# If ROUND is not given, then you receive the final table.
if(ROUND > (length(unique(games$Team.2)) - 1) * 2){
stop(paste("only ", " Rounds played in this season",
sep = as.character((length(unique(games$Team.2)) - 1) * 2)))
# Error, if too many rounds were given
}
# Data cleansing
games$Round <- as.numeric(games$Round)
games <- games[order(games$Round), ]
temp <- as.character(games$FT)
temp <- strsplit(temp, split = "-")
temp <- unlist(temp)
homegoal <- temp[c(TRUE, FALSE)]
awaygoal <- temp[c(FALSE, TRUE)]
games <- cbind(games, homegoal = as.numeric(homegoal), awaygoal = as.numeric(awaygoal))
games$FT <- NULL
# Create empty dataframe
df <- data.frame(Position = integer(length(unique(games$Team.1))),
Team = character(length(unique(games$Team.1))),
Played = integer(length(unique(games$Team.1))),
Won = integer(length(unique(games$Team.1))),
Drawn = integer(length(unique(games$Team.1))),
Lost = integer(length(unique(games$Team.1))),
GoalsFor = integer(length(unique(games$Team.1))),
GoalsAgainst = integer(length(unique(games$Team.1))),
GoalDifference = integer(length(unique(games$Team.1))),
Points = integer(length(unique(games$Team.1))))
df$Team <- unique(games$Team.1)
# for-loop to compute the points of every team
for(i in 1:(ROUND * (length(unique(games$Team.1)) / 2))){
if(games$homegoal[i] > games$awaygoal[i]){
df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 3
df$Won[games$Team.1[i] == df$Team] <- df$Won[games$Team.1[i] == df$Team] + 1
df$Lost[games$Team.2[i] == df$Team] <- df$Lost[games$Team.2[i] == df$Team] + 1
}
if(games$homegoal[i] < games$awaygoal[i]){
df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 3
df$Won[games$Team.2[i] == df$Team] <- df$Won[games$Team.2[i] == df$Team] + 1
df$Lost[games$Team.1[i] == df$Team] <- df$Lost[games$Team.1[i] == df$Team] + 1
}
if(games$homegoal[i] == games$awaygoal[i]){
(df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 1) && (df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 1)
df$Drawn[games$Team.1[i] == df$Team] <- df$Drawn[games$Team.1[i] == df$Team] + 1
df$Drawn[games$Team.2[i] == df$Team] <- df$Drawn[games$Team.2[i] == df$Team] + 1
}
}
# for-loop to compute the Goals and played games
for(i in 1:(ROUND * (length(unique(games$Team.1)) / 2))){
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
df$Played[games$Team.1[i] == df$Team] <- df$Played[games$Team.1[i] == df$Team] + 1
df$Played[games$Team.2[i] == df$Team] <- df$Played[games$Team.2[i] == df$Team] + 1
}
# Computing the Goaldifference
for(i in 1:nrow(df)){
df$GoalDifference <- df$GoalsFor - df$GoalsAgainst
}
# order the table and conditions for a tie in points
position <- order(df$Points, df$GoalDifference, df$GoalsFor, decreasing = TRUE)
df <- df[position, ]
df$Position <- 1:length(unique(games$Team.1))
rownames(df) <- NULL
return(df)
}
#------------------------------------------------------------------
# Function to pick one team to see its stats
#------------------------------------------------------------------
team <- function(t, ROUND = (length(unique(data$Team.2)) - 1) * 2){
# Error message if a wrong amount of rounds were passed
if(ROUND > (length(unique(data$Team.2)) - 1) * 2){
stop(paste("only ", " Rounds played in this season",
sep = as.character((length(unique(data$Team.2)) - 1) * 2)))
}
# Using Fuzzy Matching
temp <- unlist(lapply(t, agrep, x = unique(data$Team.1), value = TRUE, max = 1))
res <- league.table(data, ROUND)[league.table(data, ROUND)$Team == temp, ]
return(res)
}
#----------------------------------------------------------------
# Plot of table history during the whole season
#----------------------------------------------------------------
standingplot <- function(games, TEAM = NULL){
# Function that creates the standings for every team in every round
# games ... data frame with all games in one season
# TEAM ... show the table history during the season of only one team
# WARNING: It takes 30-50 seconds to create the plot
dev.new() # Opening a new plot window
# If no team was given
if(is.null(TEAM)){
# Creating an empty data frame with the team names as row names and as many columns as rounds played
a <- data.frame(row.names = unique(games$Team.1))
a[, ncol(a) + 1:((length(unique(games$Team.2)) - 1) * 2)] = NA
# Fill the empty data frame with the positions of every team in every round
for(j in 1:length(unique(games$Team.1))){
for(i in 1:((length(unique(games$Team.2)) - 1) * 2)){
a[j, i] <- league.table(games, i)$Position[league.table(games, i)$Team == rownames(a)[j]]
}
}
# Colors for the teams
colorsall <- c("red", "green", "blue", "black", "yellow", "magenta", "grey", "orange", "cyan", "seagreen",
"darkred", "darkgreen", "darkblue", "yellow4", "darkmagenta", "darkgrey", "darkcyan", "darkseagreen",
"darkorange", "beige")
# Plot the data
par(mar = c(10, 4, 5, 2), xpd = TRUE)
plot(1:((length(unique(games$Team.2)) - 1) * 2), 1:((length(unique(games$Team.2)) - 1) * 2), ylim = c(length(unique(games$Team.1)), 1),
type = "n", xlab = NA, ylab = "Position", xaxt = "n", yaxt = "n", main = "Round")
axis(3, at = 1:((length(unique(games$Team.2)) - 1) * 2), cex.axis = 0.5)
axis(2, at = 1:length(unique(games$Team.1)), cex.axis = 0.5)
legend(21, unique(games$Team.1), col = colorsall, pch = 20, ncol = 3, bty = "n")
par(xpd = FALSE)
abline(h = 1:length(unique(games$Team.1)), col = "grey")
for(i in 1:length(unique(games$Team.1))){
points(1:((length(unique(games$Team.2)) - 1) * 2), a[i, ], type = "o", col = colorsall[i],
pch = 19, cex = 0.75)
}
}
# If a particular was given
else{
temp <- rep(0, (length(unique(games$Team.2)) - 1) * 2)
par(mar = c(4, 4, 4, 4), xpd = FALSE)
# Only compute the position in every round of one team
for(i in 1:((length(unique(games$Team.2)) - 1) * 2)){
temp[i] <- league.table(games, i)$Position[league.table(games, i)$Team == TEAM]
plot(1:((length(unique(games$Team.2)) - 1) * 2), temp, type = "o", ylim = c(length(unique(games$Team.1)), 1),
pch = 19, xaxt = "n", yaxt = "n", main = paste("Position", TEAM), xlab = "Round", ylab = "Position")
axis(1, at = 1:((length(unique(games$Team.2)) - 1) * 2), cex.axis = 0.5)
axis(2, at = 1:length(unique(games$Team.1)), cex.axis = 0.5)
abline(h = 1:length(unique(games$Team.1)))
}
}
}
#----------------------------------------------------------------
# Plot Leader - Last
#----------------------------------------------------------------
lplot <- function(games){
# Gives you a visualization of the leader and the team on the last position.
# games ... data frame with all matches
# Compute the teams that are on the first position in every round
temp <- rep(0, length(unique(games$Round)))
for(i in 1:length(unique(games$Round))){
temp[i] <- league.table(games, i)$Team[league.table(games, i)$Position == 1]
}
# Edit everything to create a barplot
names(temp) <- temp
temp1 <- rep(1, length(unique(games$Round)))
names(temp1) <- names(temp)
temp1 <- as.matrix(temp1)
# Colors for the teams
colorsall <- c("red", "green", "blue", "black", "yellow", "magenta", "grey", "orange", "cyan", "seagreen",
"darkred", "darkgreen", "darkblue", "yellow4", "darkmagenta", "darkgrey", "darkcyan", "darkseagreen",
"darkorange", "beige")
# Loop that allocates the correct color to the teams of the vector created at the beginning
farben <- data.frame(Teams = unique(games$Team.1), farben = colorsall[1:length(unique(games$Team.1))])
teamfarben <- rep(0, max(unique(games$Round)))
for(i in 1:length(unique(games$Team.1))){
for(j in 1:length(unique(games$Round))){
if(any(rownames(temp1)[j] == farben$Teams[i]) == TRUE){
teamfarben[j] <- farben$farben[farben$Teams == farben$Teams[i]]
}
else{
next
}
}
}
# Plot the vector with the fitting colors
par(mar = c(1, 4, 4, 8), xpd = TRUE, mfrow = c(1, 2))
barplot(temp1, beside = FALSE, col = teamfarben, ylim = NULL, border = NA, yaxt = "n", main = "Leader")
axis(2, at = 1:max(unique(games$Round)), las = 1, cex.axis = 0.5)
legend(max(unique(games$Round)), unique(rownames(temp1)), col = unique(teamfarben), pch = 19, cex = 0.6)
# Compute the teams that are on the last position in every round
temp2 <- rep(0, length(unique(games$Round)))
for(i in 1:length(unique(games$Round))){
temp2[i] <- league.table(games, i)$Team[league.table(games, i)$Position == length(unique(games$Team.1))]
}
names(temp2) <- temp2
temp3 <- rep(1, length(unique(games$Round)))
names(temp3) <- names(temp2)
temp3 <- as.matrix(temp3)
# Colors for the teams
colorsall <- c("red", "green", "blue", "black", "yellow", "magenta", "grey", "orange", "cyan", "seagreen",
"darkred", "darkgreen", "darkblue", "yellow4", "darkmagenta", "darkgrey", "darkcyan", "darkseagreen",
"darkorange", "beige")
# Loop that allocates the correct color to the teams of the vector created at the beginning
farben <- data.frame(Teams = unique(games$Team.1), farben = colorsall[1:length(unique(games$Team.1))])
teamfarben <- rep(0, max(unique(games$Round)))
for(i in 1:length(unique(games$Team.1))){
for(j in 1:length(unique(games$Round))){
if(any(rownames(temp3)[j] == farben$Teams[i]) == TRUE){
teamfarben[j] <- farben$farben[farben$Teams == farben$Teams[i]]
}
else{
next
}
}
}
# Plot the vector with the fitting colors
barplot(temp3, beside = FALSE, col = teamfarben, ylim = NULL, border = NA, yaxt = "n", main = "Last")
axis(2, at = 1:max(unique(games$Round)), las = 1, cex.axis = 0.5)
legend(max(unique(games$Round)), unique(rownames(temp3)), col = unique(teamfarben), pch = 19, cex = 0.6)
}
# ------------------------------------
# Cross table
# ------------------------------------
crosstab <- function(games){
# Function that gives you the cross table of all matches
# games ... data frame with all matches
# Create a data frame with as many columns and rows as teams in the season
d <- data.frame(row.names = unique(games$Team.1))
d[, ncol(d) + 1:length(unique(games$Team.2))] = NA
colnames(d) <- unique(games$Team.1)
# Loops to find the matching teams and insert the result in the data frame
for(j in 1:length(colnames(d))){
for(k in 1:length(rownames(d))){
for(i in 1:length(games$Team.1)){
if(games$Team.1[i] == colnames(d)[j] & games$Team.2[i] == rownames(d)[k]){
d[colnames(d)[j], rownames(d)[k]] <- games$FT[i]
}
}
}
}
print(d)
}
#----------------------------------
# Home- and Awaytable
#----------------------------------
HAtable <- function(games, type = "Home"){
# Function to create the home and away table
# type can be "Home" and "away"
# Very similar to league.table() function
# Data cleansing
games$Round <- as.numeric(games$Round)
games <- games[order(games$Round), ]
temp <- as.character(games$FT)
temp <- strsplit(temp, split = "-")
temp <- unlist(temp)
homegoal <- temp[c(TRUE, FALSE)]
awaygoal <- temp[c(FALSE, TRUE)]
games <- cbind(games, homegoal = as.numeric(homegoal), awaygoal = as.numeric(awaygoal))
games$FT <- NULL
# Create empty data frame
df <- data.frame(Position = integer(length(unique(games$Team.1))),
Team = character(length(unique(games$Team.1))),
Played = integer(length(unique(games$Team.1))),
Won = integer(length(unique(games$Team.1))),
Drawn = integer(length(unique(games$Team.1))),
Lost = integer(length(unique(games$Team.1))),
GoalsFor = integer(length(unique(games$Team.1))),
GoalsAgainst = integer(length(unique(games$Team.1))),
GoalDifference = integer(length(unique(games$Team.1))),
Points = integer(length(unique(games$Team.1))))
df$Team <- unique(games$Team.1)
# Point of view of the away team
# Computing the points by loosing, winning or draw
if(type == "Away"){
for(i in 1:nrow(games)){
if(games$homegoal[i] < games$awaygoal[i]){
df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 3
df$Won[games$Team.2[i] == df$Team] <- df$Won[games$Team.2[i] == df$Team] + 1
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
}
if(games$homegoal[i] == games$awaygoal[i]){
df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 1
df$Drawn[games$Team.2[i] == df$Team] <- df$Drawn[games$Team.2[i] == df$Team] + 1
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
}
if(games$homegoal[i] > games$awaygoal[i]){
df$Lost[games$Team.2[i] == df$Team] <- df$Lost[games$Team.2[i] == df$Team] + 1
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
}
}
# Computing the goal difference
for(i in 1:nrow(df)){
df$GoalDifference <- df$GoalsFor - df$GoalsAgainst
}
# Order the table and conditions for a tie in points
position <- order(df$Points, df$GoalDifference, df$GoalsFor, decreasing = TRUE)
df <- df[position, ]
df$Position <- 1:length(unique(games$Team.1))
df$Played <- max(games$Round) / 2
rownames(df) <- NULL
return(df)
}
# Point of view of the home team
# Computing the points by loosing, winning or draw
else{
for(i in 1:nrow(games)){
if(games$homegoal[i] > games$awaygoal[i]){
df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 3
df$Won[games$Team.1[i] == df$Team] <- df$Won[games$Team.1[i] == df$Team] + 1
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
}
if(games$homegoal[i] == games$awaygoal[i]){
df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 1
df$Drawn[games$Team.1[i] == df$Team] <- df$Drawn[games$Team.1[i] == df$Team] + 1
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
}
if(games$homegoal[i] < games$awaygoal[i]){
df$Lost[games$Team.1[i] == df$Team] <- df$Lost[games$Team.1[i] == df$Team] + 1
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
}
}
# Computing the goal difference
for(i in 1:nrow(df)){
df$GoalDifference <- df$GoalsFor - df$GoalsAgainst
}
# Order the table and conditions for a tie in points
position <- order(df$Points, df$GoalDifference, df$GoalsFor, decreasing = TRUE)
df <- df[position, ]
df$Position <- 1:length(unique(games$Team.1))
df$Played <- max(games$Round) / 2
rownames(df) <- NULL
return(df)
}
}
#------------------------------------------------------------------------------------------
# Show relative frequencies of goals and compare with the Normal or Poisson distribution
#-----------------------------------------------------------------------------------------
Freqgoalplot <- function(games, TEAM = NULL, distribution = NULL, in.match = FALSE){
# This function shows the relative frequencies of the goals and can also compare them with
# the Normal and Poisson distribution
# games ... data frame with all games of one season
# TEAM ... show relative frequencies of a particular team
# distribution ... choose between "Poisson" and "Normal"
# in.match ... if TRUE all goals of the matches were used
# Required package
library(ggplot2)
# Data Cleansing
games$Round <- as.numeric(games$Round)
games <- games[order(games$Round), ]
temp <- as.character(games$FT)
temp <- strsplit(temp, split = "-")
temp <- unlist(temp)
homegoal <- temp[c(TRUE, FALSE)]
awaygoal <- temp[c(FALSE, TRUE)]
games <- cbind(games, homegoal = as.numeric(homegoal), awaygoal = as.numeric(awaygoal))
games$FT <- NULL
games$total <- games$homegoal + games$awaygoal
if(in.match == TRUE){
if(is.null(TEAM)){
if(is.null(distribution)){
# Create a gapless frequency table of the goals
dist <- table(c(games$total))
dist.full <- rep(0, max(games$total) + 1)
names(dist.full) <- 0:max(games$total)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Create a data frame with the frequencies and relative frequencies
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)))
rownames(data_freq) <- 0:max(games$total)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Prob)) +
geom_histogram(stat = "identity", fill = "blue") +
labs(title = "Relative frequencies of goals in one Match", x = "Goals", y = "Frequency")
} else if(isTRUE(distribution == "Poisson")){
# Create a gapless frequency table of the goals
dist <- table(c(games$total))
dist.full <- rep(0, max(games$total) + 1)
names(dist.full) <- 0:max(games$total)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Poisson distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute Poisson distribution
temp <- dpois(0:max(games$total), lambda = wmean)
poi <- data.frame(Val = temp, row.names = 0:max(games$total))
# Create a data frame with the frequencies and relative frequencies and the values of the Poisson distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Poi = as.numeric(poi$Val))
rownames(data_freq) <- 0:max(games$total)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Poi, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Poisson Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Poi"))
} else if(isTRUE(distribution == "Normal")){
# Create a gapless frequency table of the goals
dist <- table(c(games$total))
dist.full <- rep(0, max(games$total) + 1)
names(dist.full) <- 0:max(games$total)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Normal distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute the weighted standard deviation
wmean2 <- weighted.mean(as.numeric(names(dist.full))^2, w = dist.full)
sig <- sqrt(wmean2 - wmean^2)
# Compute Normal distribution
temp <- dnorm(0:max(games$total), mean = wmean, sd = sig)
norm <- data.frame(Val = temp, row.names = 0:max(games$total))
# Create a data frame with the frequencies and relative frequencies and the values of the Normal distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Norm = as.numeric(norm$Val))
rownames(data_freq) <- 0:max(games$total)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Norm, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Normal Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Norm"))
}
} else{
stop("Can't use TEAM argument, when in.match = TRUE")
}
}
else{
if(is.null(TEAM)){
if(is.null(distribution)){
# Create a gapless frequency table of the goals
dist <- table(c(games$awaygoal, games$homegoal))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Create a data frame with the frequencies and relative frequencies
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Prob)) +
geom_histogram(stat = "identity", fill = "blue") +
labs(title = "Relative frequencies of goals socred per team in one Match", x = "Goals", y = "Frequency")
} else if(isTRUE(distribution == "Normal")){
# Create a gapless frequency table of the goals
dist <- table(c(games$awaygoal, games$homegoal))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Normal distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute the weighted standard deviation for the Normal distribution
wmean2 <- weighted.mean(as.numeric(names(dist.full))^2, w = dist.full)
sig <- sqrt(wmean2 - wmean^2)
# Compute Normal distribution
temp <- dnorm(0:max(games$awaygoal, games$homegoal), mean = wmean, sd = sig)
norm <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Normal distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Norm = as.numeric(norm$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Norm, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Normal Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Norm"))
}
else if(isTRUE(distribution == "Poisson")){
# Create a gapless frequency table of the goals
dist <- table(c(games$awaygoal, games$homegoal))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Poisson distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute Poisson distribution
temp <- dpois(0:max(games$awaygoal, games$homegoal), lambda = wmean)
poi <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Poisson distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Poi = as.numeric(poi$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Poi, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Density of Goals scored per Team in each Match", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("Poi", "Mean"))
}
}
else {
if(isTRUE(distribution == "Poisson")){
# Create a gapless frequency table of the goals in relation to one particular team
dist <- table(c(games$homegoal[games$Team.1 == TEAM], games$awaygoal[games$Team.2 == TEAM]))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Poisson distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute Poisson distribution
temp <- dpois(0:max(games$awaygoal, games$homegoal), lambda = wmean)
poi <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Poisson distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Poi = as.numeric(poi$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Poi, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Poisson Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Poi"))
} else if(isTRUE(distribution == "Normal")){
# Create a gapless frequency table of the goals in relation to one particular team
dist <- table(c(games$homegoal[games$Team.1 == TEAM], games$awaygoal[games$Team.2 == TEAM]))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Normal distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute the weighted standard deviation for the Normal distribution
wmean2 <- weighted.mean(as.numeric(names(dist.full))^2, w = dist.full)
sig <- sqrt(wmean2 - wmean^2)
# Compute Normal distribution
temp <- dnorm(0:max(games$awaygoal, games$homegoal), mean = wmean, sd = sig)
norm <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Normal distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Norm = as.numeric(norm$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Norm, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = paste("Relative frequencies vs. Normal Distribution -", TEAM, sep = " "), x = "Goals", y = "Probability") +
scale_color_discrete(name = "", labels = c("RF", "Norm"))
}
else{
# Create a gapless frequency table of the goals in relation to one particular team
dist <- table(c(games$homegoal[games$Team.1 == TEAM], games$awaygoal[games$Team.2 == TEAM]))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Create a data frame with the frequencies and relative frequencies
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Prob)) +
geom_histogram(stat = "identity", fill = "blue") +
labs(title = paste("Relative frequencies of Goals:", TEAM, sep = " "), x = "Goals", y = "Frequency")
}
}
}
}
allteams <- function(games){
print(unique(games$Team.1))
}
GermanBundesliga <- read.csv("de.1.csv", encoding = "UTF-8") # German Bundesliga Season 2019/20
PremierLeague <- read.csv("eng.1.csv", encoding = "UTF-8") # Premier League Season 2017/18
PrimeraDivision <- read.csv("es.1.csv", encodin = "UTF-8") # Premiera Division Season 2017/18
Seasons <- function(){
# Shows the included data sets
print("German Bundesliga Season 2019/20 as GermanBundesliga")
print("Premier League Season 2017/18 as PremierLeague")
print("remiera Division Season 2017/18 as PrimeraDivision")
}
kosge1911/SeasonStats documentation built on Dec. 21, 2021, 7:44 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions Seasons allteams Freqgoalplot HAtable crosstab lplot standingplot team league.table allteams
Documented in allteams crosstab Freqgoalplot HAtable league.table lplot Seasons standingplot team
###################################
# Code for the SeasonStats Package
###################################
#------------------------------------------------------------
# Included data sets
#------------------------------------------------------------
GermanBundesliga <- read.csv("de.1.csv", encoding = "UTF-8")
PremierLeague <- read.csv("eng.1.csv", encoding = "UTF-8")
PrimeraDivision <- read.csv("es.1.csv", encoding = "UTF-8")
#------------------------------------------------------------
# Function that shows you all teams
#------------------------------------------------------------
allteams <- function(games){
print(unique(games$Team.1))
}
#------------------------------------------------------------
# Function for standings during and at the end of the season
#------------------------------------------------------------
league.table <- function(games, ROUND = (length(unique(games$Team.2)) - 1) * 2){
# A function that creates the table of the season.
# games ... dataframe with all games of one season
# ROUND ... shows table of specific round
# If ROUND is not given, then you receive the final table.
if(ROUND > (length(unique(games$Team.2)) - 1) * 2){
stop(paste("only ", " Rounds played in this season",
sep = as.character((length(unique(games$Team.2)) - 1) * 2)))
# Error, if too many rounds were given
}
# Data cleansing
games$Round <- as.numeric(games$Round)
games <- games[order(games$Round), ]
temp <- as.character(games$FT)
temp <- strsplit(temp, split = "-")
temp <- unlist(temp)
homegoal <- temp[c(TRUE, FALSE)]
awaygoal <- temp[c(FALSE, TRUE)]
games <- cbind(games, homegoal = as.numeric(homegoal), awaygoal = as.numeric(awaygoal))
games$FT <- NULL
# Create empty dataframe
df <- data.frame(Position = integer(length(unique(games$Team.1))),
Team = character(length(unique(games$Team.1))),
Played = integer(length(unique(games$Team.1))),
Won = integer(length(unique(games$Team.1))),
Drawn = integer(length(unique(games$Team.1))),
Lost = integer(length(unique(games$Team.1))),
GoalsFor = integer(length(unique(games$Team.1))),
GoalsAgainst = integer(length(unique(games$Team.1))),
GoalDifference = integer(length(unique(games$Team.1))),
Points = integer(length(unique(games$Team.1))))
df$Team <- unique(games$Team.1)
# for-loop to compute the points of every team
for(i in 1:(ROUND * (length(unique(games$Team.1)) / 2))){
if(games$homegoal[i] > games$awaygoal[i]){
df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 3
df$Won[games$Team.1[i] == df$Team] <- df$Won[games$Team.1[i] == df$Team] + 1
df$Lost[games$Team.2[i] == df$Team] <- df$Lost[games$Team.2[i] == df$Team] + 1
}
if(games$homegoal[i] < games$awaygoal[i]){
df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 3
df$Won[games$Team.2[i] == df$Team] <- df$Won[games$Team.2[i] == df$Team] + 1
df$Lost[games$Team.1[i] == df$Team] <- df$Lost[games$Team.1[i] == df$Team] + 1
}
if(games$homegoal[i] == games$awaygoal[i]){
(df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 1) && (df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 1)
df$Drawn[games$Team.1[i] == df$Team] <- df$Drawn[games$Team.1[i] == df$Team] + 1
df$Drawn[games$Team.2[i] == df$Team] <- df$Drawn[games$Team.2[i] == df$Team] + 1
}
}
# for-loop to compute the Goals and played games
for(i in 1:(ROUND * (length(unique(games$Team.1)) / 2))){
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
df$Played[games$Team.1[i] == df$Team] <- df$Played[games$Team.1[i] == df$Team] + 1
df$Played[games$Team.2[i] == df$Team] <- df$Played[games$Team.2[i] == df$Team] + 1
}
# Computing the Goaldifference
for(i in 1:nrow(df)){
df$GoalDifference <- df$GoalsFor - df$GoalsAgainst
}
# order the table and conditions for a tie in points
position <- order(df$Points, df$GoalDifference, df$GoalsFor, decreasing = TRUE)
df <- df[position, ]
df$Position <- 1:length(unique(games$Team.1))
rownames(df) <- NULL
return(df)
}
#------------------------------------------------------------------
# Function to pick one team to see its stats
#------------------------------------------------------------------
team <- function(t, ROUND = (length(unique(data$Team.2)) - 1) * 2){
# Error message if a wrong amount of rounds were passed
if(ROUND > (length(unique(data$Team.2)) - 1) * 2){
stop(paste("only ", " Rounds played in this season",
sep = as.character((length(unique(data$Team.2)) - 1) * 2)))
}
# Using Fuzzy Matching
temp <- unlist(lapply(t, agrep, x = unique(data$Team.1), value = TRUE, max = 1))
res <- league.table(data, ROUND)[league.table(data, ROUND)$Team == temp, ]
return(res)
}
#----------------------------------------------------------------
# Plot of table history during the whole season
#----------------------------------------------------------------
standingplot <- function(games, TEAM = NULL){
# Function that creates the standings for every team in every round
# games ... data frame with all games in one season
# TEAM ... show the table history during the season of only one team
# WARNING: It takes 30-50 seconds to create the plot
dev.new() # Opening a new plot window
# If no team was given
if(is.null(TEAM)){
# Creating an empty data frame with the team names as row names and as many columns as rounds played
a <- data.frame(row.names = unique(games$Team.1))
a[, ncol(a) + 1:((length(unique(games$Team.2)) - 1) * 2)] = NA
# Fill the empty data frame with the positions of every team in every round
for(j in 1:length(unique(games$Team.1))){
for(i in 1:((length(unique(games$Team.2)) - 1) * 2)){
a[j, i] <- league.table(games, i)$Position[league.table(games, i)$Team == rownames(a)[j]]
}
}
# Colors for the teams
colorsall <- c("red", "green", "blue", "black", "yellow", "magenta", "grey", "orange", "cyan", "seagreen",
"darkred", "darkgreen", "darkblue", "yellow4", "darkmagenta", "darkgrey", "darkcyan", "darkseagreen",
"darkorange", "beige")
# Plot the data
par(mar = c(10, 4, 5, 2), xpd = TRUE)
plot(1:((length(unique(games$Team.2)) - 1) * 2), 1:((length(unique(games$Team.2)) - 1) * 2), ylim = c(length(unique(games$Team.1)), 1),
type = "n", xlab = NA, ylab = "Position", xaxt = "n", yaxt = "n", main = "Round")
axis(3, at = 1:((length(unique(games$Team.2)) - 1) * 2), cex.axis = 0.5)
axis(2, at = 1:length(unique(games$Team.1)), cex.axis = 0.5)
legend(21, unique(games$Team.1), col = colorsall, pch = 20, ncol = 3, bty = "n")
par(xpd = FALSE)
abline(h = 1:length(unique(games$Team.1)), col = "grey")
for(i in 1:length(unique(games$Team.1))){
points(1:((length(unique(games$Team.2)) - 1) * 2), a[i, ], type = "o", col = colorsall[i],
pch = 19, cex = 0.75)
}
}
# If a particular was given
else{
temp <- rep(0, (length(unique(games$Team.2)) - 1) * 2)
par(mar = c(4, 4, 4, 4), xpd = FALSE)
# Only compute the position in every round of one team
for(i in 1:((length(unique(games$Team.2)) - 1) * 2)){
temp[i] <- league.table(games, i)$Position[league.table(games, i)$Team == TEAM]
plot(1:((length(unique(games$Team.2)) - 1) * 2), temp, type = "o", ylim = c(length(unique(games$Team.1)), 1),
pch = 19, xaxt = "n", yaxt = "n", main = paste("Position", TEAM), xlab = "Round", ylab = "Position")
axis(1, at = 1:((length(unique(games$Team.2)) - 1) * 2), cex.axis = 0.5)
axis(2, at = 1:length(unique(games$Team.1)), cex.axis = 0.5)
abline(h = 1:length(unique(games$Team.1)))
}
}
}
#----------------------------------------------------------------
# Plot Leader - Last
#----------------------------------------------------------------
lplot <- function(games){
# Gives you a visualization of the leader and the team on the last position.
# games ... data frame with all matches
# Compute the teams that are on the first position in every round
temp <- rep(0, length(unique(games$Round)))
for(i in 1:length(unique(games$Round))){
temp[i] <- league.table(games, i)$Team[league.table(games, i)$Position == 1]
}
# Edit everything to create a barplot
names(temp) <- temp
temp1 <- rep(1, length(unique(games$Round)))
names(temp1) <- names(temp)
temp1 <- as.matrix(temp1)
# Colors for the teams
colorsall <- c("red", "green", "blue", "black", "yellow", "magenta", "grey", "orange", "cyan", "seagreen",
"darkred", "darkgreen", "darkblue", "yellow4", "darkmagenta", "darkgrey", "darkcyan", "darkseagreen",
"darkorange", "beige")
# Loop that allocates the correct color to the teams of the vector created at the beginning
farben <- data.frame(Teams = unique(games$Team.1), farben = colorsall[1:length(unique(games$Team.1))])
teamfarben <- rep(0, max(unique(games$Round)))
for(i in 1:length(unique(games$Team.1))){
for(j in 1:length(unique(games$Round))){
if(any(rownames(temp1)[j] == farben$Teams[i]) == TRUE){
teamfarben[j] <- farben$farben[farben$Teams == farben$Teams[i]]
}
else{
next
}
}
}
# Plot the vector with the fitting colors
par(mar = c(1, 4, 4, 8), xpd = TRUE, mfrow = c(1, 2))
barplot(temp1, beside = FALSE, col = teamfarben, ylim = NULL, border = NA, yaxt = "n", main = "Leader")
axis(2, at = 1:max(unique(games$Round)), las = 1, cex.axis = 0.5)
legend(max(unique(games$Round)), unique(rownames(temp1)), col = unique(teamfarben), pch = 19, cex = 0.6)
# Compute the teams that are on the last position in every round
temp2 <- rep(0, length(unique(games$Round)))
for(i in 1:length(unique(games$Round))){
temp2[i] <- league.table(games, i)$Team[league.table(games, i)$Position == length(unique(games$Team.1))]
}
names(temp2) <- temp2
temp3 <- rep(1, length(unique(games$Round)))
names(temp3) <- names(temp2)
temp3 <- as.matrix(temp3)
# Colors for the teams
colorsall <- c("red", "green", "blue", "black", "yellow", "magenta", "grey", "orange", "cyan", "seagreen",
"darkred", "darkgreen", "darkblue", "yellow4", "darkmagenta", "darkgrey", "darkcyan", "darkseagreen",
"darkorange", "beige")
# Loop that allocates the correct color to the teams of the vector created at the beginning
farben <- data.frame(Teams = unique(games$Team.1), farben = colorsall[1:length(unique(games$Team.1))])
teamfarben <- rep(0, max(unique(games$Round)))
for(i in 1:length(unique(games$Team.1))){
for(j in 1:length(unique(games$Round))){
if(any(rownames(temp3)[j] == farben$Teams[i]) == TRUE){
teamfarben[j] <- farben$farben[farben$Teams == farben$Teams[i]]
}
else{
next
}
}
}
# Plot the vector with the fitting colors
barplot(temp3, beside = FALSE, col = teamfarben, ylim = NULL, border = NA, yaxt = "n", main = "Last")
axis(2, at = 1:max(unique(games$Round)), las = 1, cex.axis = 0.5)
legend(max(unique(games$Round)), unique(rownames(temp3)), col = unique(teamfarben), pch = 19, cex = 0.6)
}
# ------------------------------------
# Cross table
# ------------------------------------
crosstab <- function(games){
# Function that gives you the cross table of all matches
# games ... data frame with all matches
# Create a data frame with as many columns and rows as teams in the season
d <- data.frame(row.names = unique(games$Team.1))
d[, ncol(d) + 1:length(unique(games$Team.2))] = NA
colnames(d) <- unique(games$Team.1)
# Loops to find the matching teams and insert the result in the data frame
for(j in 1:length(colnames(d))){
for(k in 1:length(rownames(d))){
for(i in 1:length(games$Team.1)){
if(games$Team.1[i] == colnames(d)[j] & games$Team.2[i] == rownames(d)[k]){
d[colnames(d)[j], rownames(d)[k]] <- games$FT[i]
}
}
}
}
print(d)
}
#----------------------------------
# Home- and Awaytable
#----------------------------------
HAtable <- function(games, type = "Home"){
# Function to create the home and away table
# type can be "Home" and "away"
# Very similar to league.table() function
# Data cleansing
games$Round <- as.numeric(games$Round)
games <- games[order(games$Round), ]
temp <- as.character(games$FT)
temp <- strsplit(temp, split = "-")
temp <- unlist(temp)
homegoal <- temp[c(TRUE, FALSE)]
awaygoal <- temp[c(FALSE, TRUE)]
games <- cbind(games, homegoal = as.numeric(homegoal), awaygoal = as.numeric(awaygoal))
games$FT <- NULL
# Create empty data frame
df <- data.frame(Position = integer(length(unique(games$Team.1))),
Team = character(length(unique(games$Team.1))),
Played = integer(length(unique(games$Team.1))),
Won = integer(length(unique(games$Team.1))),
Drawn = integer(length(unique(games$Team.1))),
Lost = integer(length(unique(games$Team.1))),
GoalsFor = integer(length(unique(games$Team.1))),
GoalsAgainst = integer(length(unique(games$Team.1))),
GoalDifference = integer(length(unique(games$Team.1))),
Points = integer(length(unique(games$Team.1))))
df$Team <- unique(games$Team.1)
# Point of view of the away team
# Computing the points by loosing, winning or draw
if(type == "Away"){
for(i in 1:nrow(games)){
if(games$homegoal[i] < games$awaygoal[i]){
df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 3
df$Won[games$Team.2[i] == df$Team] <- df$Won[games$Team.2[i] == df$Team] + 1
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
}
if(games$homegoal[i] == games$awaygoal[i]){
df$Points[games$Team.2[i] == df$Team] <- df$Points[games$Team.2[i] == df$Team] + 1
df$Drawn[games$Team.2[i] == df$Team] <- df$Drawn[games$Team.2[i] == df$Team] + 1
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
}
if(games$homegoal[i] > games$awaygoal[i]){
df$Lost[games$Team.2[i] == df$Team] <- df$Lost[games$Team.2[i] == df$Team] + 1
df$GoalsFor[games$Team.2[i] == df$Team] <- df$GoalsFor[games$Team.2[i] == df$Team] + games$awaygoal[i]
df$GoalsAgainst[games$Team.2[i] == df$Team] <- df$GoalsAgainst[games$Team.2[i] == df$Team] + games$homegoal[i]
}
}
# Computing the goal difference
for(i in 1:nrow(df)){
df$GoalDifference <- df$GoalsFor - df$GoalsAgainst
}
# Order the table and conditions for a tie in points
position <- order(df$Points, df$GoalDifference, df$GoalsFor, decreasing = TRUE)
df <- df[position, ]
df$Position <- 1:length(unique(games$Team.1))
df$Played <- max(games$Round) / 2
rownames(df) <- NULL
return(df)
}
# Point of view of the home team
# Computing the points by loosing, winning or draw
else{
for(i in 1:nrow(games)){
if(games$homegoal[i] > games$awaygoal[i]){
df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 3
df$Won[games$Team.1[i] == df$Team] <- df$Won[games$Team.1[i] == df$Team] + 1
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
}
if(games$homegoal[i] == games$awaygoal[i]){
df$Points[games$Team.1[i] == df$Team] <- df$Points[games$Team.1[i] == df$Team] + 1
df$Drawn[games$Team.1[i] == df$Team] <- df$Drawn[games$Team.1[i] == df$Team] + 1
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
}
if(games$homegoal[i] < games$awaygoal[i]){
df$Lost[games$Team.1[i] == df$Team] <- df$Lost[games$Team.1[i] == df$Team] + 1
df$GoalsFor[games$Team.1[i] == df$Team] <- df$GoalsFor[games$Team.1[i] == df$Team] + games$homegoal[i]
df$GoalsAgainst[games$Team.1[i] == df$Team] <- df$GoalsAgainst[games$Team.1[i] == df$Team] + games$awaygoal[i]
}
}
# Computing the goal difference
for(i in 1:nrow(df)){
df$GoalDifference <- df$GoalsFor - df$GoalsAgainst
}
# Order the table and conditions for a tie in points
position <- order(df$Points, df$GoalDifference, df$GoalsFor, decreasing = TRUE)
df <- df[position, ]
df$Position <- 1:length(unique(games$Team.1))
df$Played <- max(games$Round) / 2
rownames(df) <- NULL
return(df)
}
}
#------------------------------------------------------------------------------------------
# Show relative frequencies of goals and compare with the Normal or Poisson distribution
#-----------------------------------------------------------------------------------------
Freqgoalplot <- function(games, TEAM = NULL, distribution = NULL, in.match = FALSE){
# This function shows the relative frequencies of the goals and can also compare them with
# the Normal and Poisson distribution
# games ... data frame with all games of one season
# TEAM ... show relative frequencies of a particular team
# distribution ... choose between "Poisson" and "Normal"
# in.match ... if TRUE all goals of the matches were used
# Required package
library(ggplot2)
# Data Cleansing
games$Round <- as.numeric(games$Round)
games <- games[order(games$Round), ]
temp <- as.character(games$FT)
temp <- strsplit(temp, split = "-")
temp <- unlist(temp)
homegoal <- temp[c(TRUE, FALSE)]
awaygoal <- temp[c(FALSE, TRUE)]
games <- cbind(games, homegoal = as.numeric(homegoal), awaygoal = as.numeric(awaygoal))
games$FT <- NULL
games$total <- games$homegoal + games$awaygoal
if(in.match == TRUE){
if(is.null(TEAM)){
if(is.null(distribution)){
# Create a gapless frequency table of the goals
dist <- table(c(games$total))
dist.full <- rep(0, max(games$total) + 1)
names(dist.full) <- 0:max(games$total)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Create a data frame with the frequencies and relative frequencies
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)))
rownames(data_freq) <- 0:max(games$total)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Prob)) +
geom_histogram(stat = "identity", fill = "blue") +
labs(title = "Relative frequencies of goals in one Match", x = "Goals", y = "Frequency")
} else if(isTRUE(distribution == "Poisson")){
# Create a gapless frequency table of the goals
dist <- table(c(games$total))
dist.full <- rep(0, max(games$total) + 1)
names(dist.full) <- 0:max(games$total)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Poisson distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute Poisson distribution
temp <- dpois(0:max(games$total), lambda = wmean)
poi <- data.frame(Val = temp, row.names = 0:max(games$total))
# Create a data frame with the frequencies and relative frequencies and the values of the Poisson distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Poi = as.numeric(poi$Val))
rownames(data_freq) <- 0:max(games$total)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Poi, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Poisson Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Poi"))
} else if(isTRUE(distribution == "Normal")){
# Create a gapless frequency table of the goals
dist <- table(c(games$total))
dist.full <- rep(0, max(games$total) + 1)
names(dist.full) <- 0:max(games$total)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Normal distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute the weighted standard deviation
wmean2 <- weighted.mean(as.numeric(names(dist.full))^2, w = dist.full)
sig <- sqrt(wmean2 - wmean^2)
# Compute Normal distribution
temp <- dnorm(0:max(games$total), mean = wmean, sd = sig)
norm <- data.frame(Val = temp, row.names = 0:max(games$total))
# Create a data frame with the frequencies and relative frequencies and the values of the Normal distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Norm = as.numeric(norm$Val))
rownames(data_freq) <- 0:max(games$total)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Norm, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Normal Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Norm"))
}
} else{
stop("Can't use TEAM argument, when in.match = TRUE")
}
}
else{
if(is.null(TEAM)){
if(is.null(distribution)){
# Create a gapless frequency table of the goals
dist <- table(c(games$awaygoal, games$homegoal))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Create a data frame with the frequencies and relative frequencies
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Prob)) +
geom_histogram(stat = "identity", fill = "blue") +
labs(title = "Relative frequencies of goals socred per team in one Match", x = "Goals", y = "Frequency")
} else if(isTRUE(distribution == "Normal")){
# Create a gapless frequency table of the goals
dist <- table(c(games$awaygoal, games$homegoal))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Normal distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute the weighted standard deviation for the Normal distribution
wmean2 <- weighted.mean(as.numeric(names(dist.full))^2, w = dist.full)
sig <- sqrt(wmean2 - wmean^2)
# Compute Normal distribution
temp <- dnorm(0:max(games$awaygoal, games$homegoal), mean = wmean, sd = sig)
norm <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Normal distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Norm = as.numeric(norm$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Norm, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Normal Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Norm"))
}
else if(isTRUE(distribution == "Poisson")){
# Create a gapless frequency table of the goals
dist <- table(c(games$awaygoal, games$homegoal))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Poisson distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute Poisson distribution
temp <- dpois(0:max(games$awaygoal, games$homegoal), lambda = wmean)
poi <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Poisson distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Poi = as.numeric(poi$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Poi, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Density of Goals scored per Team in each Match", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("Poi", "Mean"))
}
}
else {
if(isTRUE(distribution == "Poisson")){
# Create a gapless frequency table of the goals in relation to one particular team
dist <- table(c(games$homegoal[games$Team.1 == TEAM], games$awaygoal[games$Team.2 == TEAM]))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Poisson distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute Poisson distribution
temp <- dpois(0:max(games$awaygoal, games$homegoal), lambda = wmean)
poi <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Poisson distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Poi = as.numeric(poi$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Poi, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = "Relative freqencies vs. Poisson Distribution", x = "Goals", y = "Frequency") +
scale_color_discrete(name = "", labels = c("RF", "Poi"))
} else if(isTRUE(distribution == "Normal")){
# Create a gapless frequency table of the goals in relation to one particular team
dist <- table(c(games$homegoal[games$Team.1 == TEAM], games$awaygoal[games$Team.2 == TEAM]))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Compute the weighted mean for the Normal distribution
wmean <- weighted.mean(as.numeric(names(dist.full)), w = dist.full)
# Compute the weighted standard deviation for the Normal distribution
wmean2 <- weighted.mean(as.numeric(names(dist.full))^2, w = dist.full)
sig <- sqrt(wmean2 - wmean^2)
# Compute Normal distribution
temp <- dnorm(0:max(games$awaygoal, games$homegoal), mean = wmean, sd = sig)
norm <- data.frame(Val = temp, row.names = 0:max(games$awaygoal, games$homegoal))
# Create a data frame with the frequencies and relative frequencies and the values of the Normal distribution
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)),
Norm = as.numeric(norm$Val))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Norm, group = 1, color = "red"), size = 1) +
geom_line() +
geom_line(aes(y = Prob, color = "blue"), size = 1) +
theme_bw() +
labs(title = paste("Relative frequencies vs. Normal Distribution -", TEAM, sep = " "), x = "Goals", y = "Probability") +
scale_color_discrete(name = "", labels = c("RF", "Norm"))
}
else{
# Create a gapless frequency table of the goals in relation to one particular team
dist <- table(c(games$homegoal[games$Team.1 == TEAM], games$awaygoal[games$Team.2 == TEAM]))
dist.full <- rep(0, max(games$awaygoal, games$homegoal) + 1)
names(dist.full) <- 0:max(games$awaygoal, games$homegoal)
dist.full[names(dist)] <- dist
cumsum_dist <- cumsum(dist.full)
# Create a data frame with the frequencies and relative frequencies
data_freq <- data.frame(Freq = as.numeric(dist.full),
Prob = as.numeric(dist.full / max(cumsum_dist)),
FreqCS = as.numeric(cumsum_dist),
ProbCS = as.numeric(cumsum_dist / max(cumsum_dist)))
rownames(data_freq) <- 0:max(games$awaygoal, games$homegoal)
# Plot the data frame
ggplot(data_freq, aes(x = rownames(data_freq), y = Prob)) +
geom_histogram(stat = "identity", fill = "blue") +
labs(title = paste("Relative frequencies of Goals:", TEAM, sep = " "), x = "Goals", y = "Frequency")
}
}
}
}
allteams <- function(games){
print(unique(games$Team.1))
}
GermanBundesliga <- read.csv("de.1.csv", encoding = "UTF-8") # German Bundesliga Season 2019/20
PremierLeague <- read.csv("eng.1.csv", encoding = "UTF-8") # Premier League Season 2017/18
PrimeraDivision <- read.csv("es.1.csv", encodin = "UTF-8") # Premiera Division Season 2017/18
Seasons <- function(){
# Shows the included data sets
print("German Bundesliga Season 2019/20 as GermanBundesliga")
print("Premier League Season 2017/18 as PremierLeague")
print("remiera Division Season 2017/18 as PrimeraDivision")
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.