FinalAnalisis.R

#LOADING LIBRARIES----------------------------------------------------------------------------------------------
library(readxl)
library(dplyr)
library(CCA)
library( CCP )
library( ggplot2 )
library( GGally )

#LOADING DATA--------------------------------------------------------------------------------------------------
data <- read_excel("world_cup_results.xlsx")

colnames(data)


dataCCA <- data[,c(1,7,8,9,10,11)]



dataPrueba2 <- dataCCA  %>% group_by(Year)


length(which(dataPrueba2$Year==2014))
  
dataPrueba25 <- summarize(dataPrueba, numPartidos = count())


dataPrueba3 <- filter(dataCCA, HomeTeam =="Brazil" | AwayTeam=="Brazil" )
dataPrueba3


dataTest <- summarise(dataPrueba2, HomeGoals = sum(HomeGoals),  AwayGoals = sum(AwayGoals))
dataTest


JuegosMundial <- matrix(c(1930,1934,1938,1950,1954,1958,1962,1966,1970,1974,1978,1982,1986,1990,1994,1998,2002,2004,2006,2010,2014,18,17,18,22,26,35,32,32,32,38,38,52,52,52,52,64,64,64,64),20,2,byrow = F)

JuegosMundial

#GOALS BY COUNTRIES---------------------------------------------------------------------------------
#HOME GOALS 
Ecuador_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Ecuador", ] )
Brazil_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Brazil", ] )
Uruguay_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Uruguay", ] )
Germany_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Germany", ] )
Italy_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Italy", ] )
France_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "France", ] )
Colombia_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Colombia", ] )
Argentina_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Argentina", ] )
Netherlands_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Netherlands", ] )
Spain_H <- as.data.frame( dataCCA[ dataCCA$HomeTeam == "Spain", ] )
#AWAY GOALS 
Ecuador_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Ecuador", ] )
Brazil_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Brazil", ] )
Uruguay_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Uruguay", ] )
Germany_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Germany", ] )
Italy_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Italy", ] )
France_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "France", ] )
Colombia_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Colombia", ] )
Argentina_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Argentina", ] )
Netherlands_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Netherlands", ] )
Spain_A <- as.data.frame( dataCCA[ dataCCA$AwayTeam == "Spain", ] )

names( Italy_A ) <- c("Year", "Country",   "HomeTeam",  "HomeGoals", "AwayGoals", "AwayTeam")
Italy_A
Ecuador_H <- select( Ecuador_H, Year, Country, HomeGoals, AwayGoals )
names( Ecuador_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Ecuador_H
Ecuador_A <-Ecuador_A[ , c( 1, 2, 5, 4 ) ]
names( Ecuador_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Ecuador_A

#ECUADOR GOALS--------------------------------------------------------------------------------------
Ec_Goals <- rbind( Ecuador_H, Ecuador_A )
Ec_Goals

#---------------------------------------------------------------------------------------------------
Brazil_H <- select( Brazil_H, Year, Country, HomeGoals, AwayGoals )
names( Brazil_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Brazil_H
Brazil_A <- Brazil_A[ , c( 1, 2, 5, 4 ) ]
names( Brazil_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Brazil_A
#BRAZIL GOALS---------------------------------------------------------------------------------------
Br_Goals <- rbind( Brazil_H, Brazil_A )
Br_Goals
#---------------------------------------------------------------------------------------------------
Germany_H <- select( Germany_H, Year, Country, HomeGoals, AwayGoals )
names( Germany_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Germany_H
Germany_A <- select( Germany_A, Year, Country, AwayGoals, HomeGoals )
names( Germany_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Brazil_A
#GERMANY GOALS--------------------------------------------------------------------------------------
Ger_Goals <- rbind( Germany_H, Germany_A )
Ger_Goals
#---------------------------------------------------------------------------------------------------
Italy_H <- select( Italy_H, Year, Country, HomeGoals, AwayGoals )
names( Italy_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Italy_H
Italy_A <- Italy_A[, c(1, 2 , 5, 4 ) ]
names( Italy_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Italy_A
#ITALY GOALS----------------------------------------------------------------------------------------
It_Goals <- rbind( Italy_H, Italy_A )
It_Goals
#---------------------------------------------------------------------------------------------------
Uruguay_H <- select( Uruguay_H, Year, Country, HomeGoals, AwayGoals )
names( Uruguay_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Uruguay_H
Uruguay_A <- select( Uruguay_A, Year, Country, AwayGoals, HomeGoals )
names( Uruguay_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Uruguay_A

names( Argentina_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )

#URUGUAY GOALS--------------------------------------------------------------------------------------
Ur_Goals <- rbind( Uruguay_H, Uruguay_A )
Ur_Goals

#---------------------------------------------------------------------------------------------------
France_H <- select( France_H, Year, Country, HomeGoals, AwayGoals )
names( France_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
France_H
France_A <- select( France_A, Year, Country, AwayGoals, HomeGoals )
names( France_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
France_A
#GERMANY GOALS--------------------------------------------------------------------------------------
Fra_Goals <- rbind( France_H, France_A )
Fra_Goals

#---------------------------------------------------------------------------------------------------
Colombia_H <- select( Colombia_H, Year, Country, HomeGoals, AwayGoals )
names( Colombia_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Colombia_H
Colombia_A <- select( Colombia_A, Year, Country, AwayGoals, HomeGoals )
names( Colombia_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Colombia_A
#COLOMBIA GOALS--------------------------------------------------------------------------------------
Col_Goals <- rbind( Colombia_H, Colombia_A )
Col_Goals

#---------------------------------------------------------------------------------------------------
Argentina_H <- select( Argentina_H, Year, Country, HomeGoals, AwayGoals )
names( Argentina_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Argentina_H
Argentina_A <- Argentina_A[ ,c( 1, 2, 5, 4 ) ]
names( Argentina_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Argentina_A
#ARGENTINA GOALS------------------------------------------------------------------------------------
Arg_Goals <- rbind( Argentina_H, Argentina_A )
Arg_Goals

#---------------------------------------------------------------------------------------------------
Netherlands_H <- select( Netherlands_H, Year, Country, HomeGoals, AwayGoals )
names( Netherlands_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Netherlands_H
Netherlands_A <- Netherlands_A[ ,c( 1, 2, 5, 4 ) ]
names( Netherlands_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Netherlands_A
#ARGENTINA GOALS------------------------------------------------------------------------------------
Neth_Goals <- rbind( Netherlands_H, Netherlands_A )
Neth_Goals

#---------------------------------------------------------------------------------------------------
Spain_H <- select( Spain_H, Year, Country, HomeGoals, AwayGoals )
names( Spain_H ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Spain_H
Spain_A <- Spain_A[ ,c( 1, 2, 5, 4 ) ]
names( Spain_A ) <- c( "Year", "Country", "GoalsFavor", "GoalsContra" )
Spain_A
#ARGENTINA GOALS------------------------------------------------------------------------------------
Spain_Goals <- rbind( Spain_H, Spain_A )
Spain_Goals

#FINAL GOALS----------------------------------------------------------------------------------------
Goals <- matrix( 0, 2, 10 )
Goals[ 1, 1 ] <- sum( Ec_Goals[ 1:4, 3 ] )
Goals[ 2, 1 ] <- sum( Ec_Goals[ 5:10, 3 ] )
Goals[ 1, 2 ] <- sum( Br_Goals[ 1:82, 3 ] )
Goals[ 2, 2 ] <- sum( Br_Goals[ 83:108, 3 ] )
Goals[ 1, 3 ] <- sum( Ger_Goals[ 1:32, 3 ] )
Goals[ 2, 3 ] <- sum( Ger_Goals[ 33:48, 3 ] )
Goals[ 1, 4 ] <- sum( It_Goals[ 1:57, 3] )
Goals[ 2, 4 ] <- sum( It_Goals[ 58:83, 3] )
Goals[ 1, 5 ] <- sum( Ur_Goals[ 1:28, 3 ] )
Goals[ 2, 5 ] <- sum( Ur_Goals[ 29:52, 3 ] )
Goals[ 1, 6 ] <- sum( Fra_Goals[ 1:31, 3 ] )
Goals[ 2, 6 ] <- sum( Fra_Goals[ 32:62, 3 ] )
Goals[ 1, 7 ] <- sum( Col_Goals[ 1:7, 3 ] )
Goals[ 2, 7 ] <- sum( Col_Goals[ 8:20, 3 ] )
Goals[ 1, 8 ] <- sum( Arg_Goals[ 1:54, 3 ] )
Goals[ 2, 8 ] <- sum( Arg_Goals[ 55:81, 3 ] )
Goals[ 1, 9 ] <- sum( Neth_Goals[ 1:32, 3 ] )
Goals[ 2, 9 ] <- sum( Neth_Goals[ 33:54, 3 ] )
Goals[ 1, 10 ] <- sum( Spain_Goals[ 1:29, 3 ] )
Goals[ 2, 10 ] <- sum( Spain_Goals[ 30:58, 3 ] )

colnames( Goals ) <- c( "Ecuador", "Brazil", "Germany", "Italy", "Uruguay", "France", "Colombia", 
                        "Argentina", "Netherlands", "Spain" )
rownames( Goals ) <- c( "Home Favor Goals", "Away Favor Goals" )
Goals

GoalsContra <- matrix( 0, 2, 10 )
GoalsContra[ 1, 1 ] <- sum( Ec_Goals[ 1:4, 4 ] )
GoalsContra[ 2, 1 ] <- sum( Ec_Goals[ 5:10, 4 ] )
GoalsContra[ 1, 2 ] <- sum( Br_Goals[ 1:82, 4 ] )
GoalsContra[ 2, 2 ] <- sum( Br_Goals[ 83:108, 4 ] )
GoalsContra[ 1, 3 ] <- sum( Ger_Goals[ 1:32, 4 ] )
GoalsContra[ 2, 3 ] <- sum( Ger_Goals[ 33:48, 4 ] )
GoalsContra[ 1, 4 ] <- sum( It_Goals[ 1:57, 4 ] )
GoalsContra[ 2, 4 ] <- sum( It_Goals[ 58:83, 4 ] )
GoalsContra[ 1, 5 ] <- sum( Ur_Goals[ 1:28, 4 ] )
GoalsContra[ 2, 5 ] <- sum( Ur_Goals[ 29:52, 4 ] )
GoalsContra[ 1, 6 ] <- sum( Fra_Goals[ 1:31, 4 ] )
GoalsContra[ 2, 6 ] <- sum( Fra_Goals[ 32:62, 4 ] )
GoalsContra[ 1, 7 ] <- sum( Col_Goals[ 1:7, 4 ] )
GoalsContra[ 2, 7 ] <- sum( Col_Goals[ 8:20, 4 ] )
GoalsContra[ 1, 8 ] <- sum( Arg_Goals[ 1:54, 4 ] )
GoalsContra[ 2, 8 ] <- sum( Arg_Goals[ 55:81, 4 ] )
GoalsContra[ 1, 9 ] <- sum( Neth_Goals[ 1:32, 4 ] )
GoalsContra[ 2, 9 ] <- sum( Neth_Goals[ 33:54, 4 ] )
GoalsContra[ 1, 10 ] <- sum( Spain_Goals[ 1:29, 4 ] )
GoalsContra[ 2, 10 ] <- sum( Spain_Goals[ 30:58, 4 ] )

colnames( GoalsContra ) <- c( "Ecuador", "Brazil", "Germany", "Italy", "Uruguay", "France", "Colombia", 
                              "Argentina", "Netherlands", "Spain" )
rownames( GoalsContra ) <- c( "Home Against Goals", "Away Against Goals" )
GoalsContra 
#SEDE MUNDIAL---------------------------------------------------------------------------------------

Sede <- matrix( c( 0, 2, 2, 2, 1, 2, 0, 1, 0, 1 ), 1, 10 )
colnames( Sede ) <- c( "Ecuador", "Brazil", "Germany", "Italy", "Uruguay", "France", "Colombia", 
                       "Argentina", "Netherlands", "Spain" )
rownames( Sede ) <- "Sede"
Sede

#PARTICIPATION--------------------------------------------------------------------------------------
participation <- matrix( c( 3, 20, 18, 18, 12, 14, 5, 16, 10, 14 ), 1, 10 )
colnames( participation ) <- c( "Ecuador", "Brazil", "Germany", "Italy", "Uruguay",  "France", "Colombia", 
                                "Argentina", "Netherlands", "Spain" )
rownames( participation ) <- "Participation"
participation

Goals <- as.data.frame( Goals )
GoalsContra <- as.data.frame( GoalsContra )
Sede <- as.data.frame( Sede )
participation <- as.data.frame( participation )

Table <- rbind( Goals, GoalsContra )
Table <- rbind( Table, Sede )
Table1 <- rbind( Table, participation )
Table1

#CANONICAL CORRELATION------------------------------------------------------------------------------
G1 <- Table1[ , 1:4 ]
G2 <- Table1[ , c( 5, 6 ) ]
matcor( G1, G2 )

cc1 <- cc( G1, G2 )
# display the canonical correlations
cc1$cor
#raw canonical coefficients
cc1[ 3:4 ]
# compute canonical loadings
cc2 <- comput( G1, G2, cc1 )

# display canonical loadings
cc2[ 3:6 ]

# tests of canonical dimensions
rho <- cc1$cor
## Define number of observations, number of variables in first set, and number of variables in the second set.
n <- dim( G1 )[ 1 ]
p <- ncol( G1 )
q <- ncol( G2 )

## Calculate p-values using the F-approximations of different test statistics:
Wilks <- p.asym(rho, n, p, q, tstat = "Wilks")
p.asym(rho, n, p, q, tstat = "Hotelling")

Tab1 <- matrix( 0, 2, 5 )
Tab1[ 1:2, 1 ] <- t( cor )
Tab1[ 1:2, 2 ] <- Wilks$approx
Tab1[ 1:2, 3 ] <- Wilks$df1
Tab1[ 1:2, 4 ] <- Wilks$df2
Tab1[ 1:2, 5 ] <- Wilks$p.value
colnames( Tab1 ) <- c( "Corr", "F", "df1", "df2", "p-value" )
rownames( Tab1 ) <- c( "dim 1", "dim 2" )
#Table 1: Tests of Canonical Dimensions-------------------------------------------------------------
Tab1
#Table 2: Standardized Canonical Coefficients-------------------------------------------------------



# Test of significance
val <- 1- sum( cc1$cor )
Chi_test <- -( n -1/2 * ( p + q + 3 ) ) * sum( log( 1 -cc1$cor ) )
Chi_test 
#15.96467

#Theorical p-value with p*q df and alpha=0.05; 15,5073 < 15,963 so, we cannot reject Ho 
#at least one of the r canonical correlations is significant.

#GLOBAL DATA CANONICAL CORRELATION ANALYSIS (CCA)---------------------------------------------------

Data2 <- read.csv( "DatosGlobales.csv", header = T )
rownames( Data2 ) <- Data2[ , 2 ]
Data2
#CHECKING NORMAL DISTRIBUTION OF THE DATA-----------------------------------------------------------
Data2 <- as.matrix( Data2 )
mvShapiro.Test( Data2 )
#p-value = 0.01748 we cannot reject Ho for alpha=0.01. Therefore, the data is normally distributed

Group1 <- Data2[ , c( 3, 5 ) ]
Group1
Group2 <- Data2[ , c( 4, 6 ) ]
colnames( Group2 ) <- c( "No. of Participants", "Attendees")
Group2

matcor( Group1, Group2 )
cc_data2 <- cc( Group1, Group2 )
# display the canonical correlations
cor2 <- cc_data2$cor
#raw canonical coefficients
cc_data2[ 3:4 ]
# compute canonical loadings
cc2_data2 <- comput( Group1, Group2, cc_data2 )
# display canonical loadings
cc2_data2[ 3:6 ]
# tests of canonical dimensions
rho <- cc_data2$cor
## Define number of observations, number of variables in first set, and number of variables in the second set.
n2 <- dim( Group1 )[ 1 ]
p2 <- ncol( Group1 )
q2 <- ncol( Group2 )

## Calculate p-values using the F-approximations of different test statistics:
Wilks2 <- p.asym(rho, n2, p2, q2, tstat = "Wilks")
p.asym(rho, n2, p2, q2, tstat = "Hotelling")

Final_tab <- matrix( 0, 2, 5 )
Final_tab[ 1:2, 1 ] <- t( cor2 )
Final_tab[ 1:2, 2 ] <- Wilks2$approx
Final_tab[ 1:2, 3 ] <- Wilks2$df1
Final_tab[ 1:2, 4 ] <- Wilks2$df2
Final_tab[ 1:2, 5 ] <- Wilks2$p.value
colnames( Final_tab ) <- c( "Corr", "F", "df1", "df2", "p-value" )
rownames( Final_tab ) <- c( "dim 1", "dim 2" )
#Table 1: Tests of Canonical Dimensions-------------------------------------------------------------
Final_tab
# Test of significance------------------------------------------------------------------------------
Chi_test2 <- -( n2 -1/2 * ( p2 + q2 + 3 ) ) * sum( log( 1 -cc_data2$cor ) )
Chi_test2
#=64.57317
#p*q degrees of freedom, alpha= 0.05, CHI_Therorical = 9.4877 which is smaller than our Chi value test 
#Therefore, we cannot reject Ho. Then at least one of the r canonical correlations is significant.