###Pearson's Data and Bivariate Boxplots

###Reading in Pearson's data from a file in .csv format.

Heights<-read.csv("Pearson's Height Data.csv")


#Calculating the univariate boxplots.
boxplot(Heights$Father)
boxplot(Heights$Son, pch=19)



###Activating the MVA library
library("MVA")


###Constructing the bivariate boxplot. See comments below.
with(Heights,bvbox(cbind(Father, Son),7, pch=20, xlab="Heights of Fathers",ylab="Heights of Sons"))

#The function, "with()", tells R the data structure in which it can find
#the data referenced in the subsequent function call. 
#This makes it easier to compose and make sense of the subsequent function call.
#The second argument for the "bvbox" function (7 in this case) specifies 
#how much larger the "fence" ellipse should be vs. the "hinge" ellipse. 
#With a value of 7, this will tend to create a fence that contains about 99%
#of a dataset that has been generated by a bivariate normal distribution. 
#The argument, "pch=20" specifies that the character to be used to represent
#points is a small circle. You can find other options in the R documentation.
#The final two arguments just specify axis labels. Note, however, that the 
#y-axis label will likely be hidden. This glitch in the bvbox function
#needs to be fixed. 


###Some script for creating three-dimensional scatterplots. 

library("scatterplot3d")

scatterplot3d(MLB$Earned.Run.Avg,MLB$Errors,MLB$Pct.Wins,type="h")