rm(list=ls())

data = read.csv("ILE-2013.csv",header=TRUE)

namevar = c("score","property","corrupt","fiscal","govern","busin","labor",
"monet","trade","invest","finan")

attach(data)

n = nrow(data)

indicadores = data[,3:13]

boxplot(indicadores,names=namevar)

pairs(indicadores,labels=namevar)

round(cor(indicadores),2)

par(mfrow=c(1,2))
plot(indicadores[,2],indicadores[,3],xlab=namevar[2],ylab=namevar[3],
xlim=c(0,100),ylim=c(0,100),pch=16)
plot(indicadores[,10],indicadores[,11],xlab=namevar[10],ylab=namevar[11],
xlim=c(0,100),ylim=c(0,100),pch=16)


continente = unique(Continente)
cont = rep(1,n)
cont[Continente=="America"]=2
cont[Continente=="Asia"]=3
cont[Continente=="Europa"]=4
cont[Continente=="Oceania"]=5


par(mfrow=c(1,2))
plot(indicadores[,2]+0.5*rnorm(176),indicadores[,3]+0.5*rnorm(176),xlab=namevar[2],ylab=namevar[3],
xlim=c(0,100),ylim=c(0,100),col=cont,pch=16)
legend(10,100,legend=continente,col=1:5,bty="n",pch=16)

plot(indicadores[,10]+0.5*rnorm(176),indicadores[,11]+0.5*rnorm(176),xlab=namevar[10],ylab=namevar[11],
xlim=c(0,100),ylim=c(0,100),col=cont,pch=16)




# calculando distancias de Mahalanobis
# ------------------------------------
x    = as.matrix(data[,4:13])
n    = nrow(x)
xbar = apply(x,2,mean)
S    = var(x)
iS   = solve(S)  
dist = x-matrix(xbar,n,10,byrow=TRUE)
iS   = matrix(as.double(iS),10,10)
Dm   = diag(dist%*%iS%*%t(dist))
pval = round(1-pchisq(Dm,3),4)
rDm  = round(sqrt(Dm),2)

par(mfrow=c(1,1))
plot(Dm,xlab="Observation",ylab="Mahalanobis distance")
abline(h=qchisq(0.95,3),col=2,lwd=2)
abline(h=qchisq(0.99,3),col=4,lwd=2)
abline(h=qchisq(0.999,3),col=6,lwd=2)
legend(100,80,c("Cut-off for 5% level","Cut-off for 1% level","Cut-off for 0.1% level"),
       col=c(2,4,6),bty="n",lwd=2)