#### Quant 1
#### Week 3 Code

require(foreign)
dail2002 <- read.dta("../02 Univariate/dail2002.dta")
dail2002$incumbf <- factor(dail2002$incumb,
                           labels=c("Challenger", "Incumbent"))

## Candidates by party
table(dail2002$party)
options("digits"=3)   # set limit on digits to print
prop.table(table(dail2002$party))  # proportions
temptable <- cbind(table(dail2002$party), prop.table(table(dail2002$party))*100)  # make into single table
colnames(temptable) <- c("N", "%")
temptable

## two-way tables
partyinctable <- table(dail2002$party, dail2002$incumbf)
addmargins(partyinctable)     # both row and column marginals
addmargins(partyinctable, 2)  # row marginals
addmargins(partyinctable, 1)  # column marginals
options("digits"=1)
prop.table(partyinctable)      # proportion of all cells
prop.table(partyinctable, 1)   # row proportions
prop.table(partyinctable, 2)   # column proportions
addmargins(prop.table(partyinctable))        # total proportion marginals
addmargins(prop.table(partyinctable, 1), 2)  # row proportions and sum
addmargins(prop.table(partyinctable, 2), 1)  # column proportions and sum

## three-way tables
(partyincgendertable <- table(dail2002$party, dail2002$incumbf, dail2002$gender))
ftable(partyincgendertable)

## barplots to compare 2 variables
barplot(partyinctable)
barplot(table(dail2002$party, dail2002$incumbf), beside=T,
        legend.text=T)
barplot(table(dail2002$incumbf, dail2002$party), beside=T, legend.text=T)

## comparative boxplot
boxplot(dail2002$votes1st ~ dail2002$party, ylab="First Preference Votes")

## compare two density plots
plot(density(subset(dail2002$votes1st, dail2002$incumbf=="Challenger"), na.rm=T),
     main="", xlab="First Preference Votes")
lines(density(subset(dail2002$votes1st, dail2002$incumbf=="Incumbent"), na.rm=T))
text(3000,.00020,"Challengers")
text(10000,.00015,"Incumbents")

## scatterplot
plot(dail2002$votes1st ~ dail2002$spend_total, xlab="Spending in Euros", ylab="First Preference Votes")

## scatterplot distinguishing challengers and incumbents
plot(subset(dail2002$votes1st, dail2002$incumbf=="Challenger") ~
     subset(dail2002$spend_total, dail2002$incumbf=="Challenger"),
     col="red",
     xlab="Spending in Euros", ylab="First Preference Votes")
points(subset(dail2002$votes1st, dail2002$incumbf=="Incumbent") ~
     subset(dail2002$spend_total, dail2002$incumbf=="Incumbent"),
     col="blue")
legend(0, 9000, c("Challenger", "Incumbent"), col=c("red","blue"), pch="o")       


sum((x-mean(x))*(y-mean(y))) /
  sqrt(sum((x-mean(x))^2)*sum((y-mean(y))^2))