# Applied Econometrics with R
# Kleiber, Christian, Zeileis, Achim 2008
# http://www.springer.com/978-0-387-77316-2

# Install packages
install.packages("AER")

# The demand for economics journals
data("Journals", package = "AER")
dim(Journals)
names(Journals)
head(Journals)
plot(log(subs) ~ log(price/citations), data = Journals)
j_lm <- lm(log(subs) ~ log(price/citations), data = Journals)
summary(j_lm)
abline(j_lm)

# regression
Journals$citeprice = Journals$price/Journals$citations
jour_lm  = lm(log(subs) ~ log(citeprice), data = Journals)
jour_slm = summary(jour_lm)
jour_slm$coefficients

# point and interval estimates
jour_lm$coefficients
confint(jour_lm, level = 0.99)

# linear hypothesis
library(car)
linear.hypothesis(jour_lm, "log(citeprice) = -.5")

# multiple regression
# Determinants of wages
data("CPS1985", package = "AER")
cps <- CPS1985
library("quantreg")
cps_lm <- lm(log(wage) ~ experience + I(experience^2) + education, data = cps)
cps_rq <- rq(log(wage) ~ experience + I(experience^2) + education, data = cps, tau = seq(0.2, 0.8, by = 0.15))
summary(cps_lm)
summary(cps_rq)
plot(summary(cps_rq))

# R
ls()
rm(qq)
?lm

#
plot(1:20, pch = 1:20, col = 1:20, cex = 2)

curve(dnorm, from = -5, to = 5, col = "slategray", lwd = 3, main = "Density of the standard normal distribution")
text(-5, 0.3, expression(f(x) == frac(1, sigma ~~ sqrt(2*pi)) ~~ e^{-frac((x - mu)^2, 2*sigma^2)}), adj = 0)

curve(x^2)

# Data
data("CPS1985", package = "AER")
head(CPS1985)

attach(CPS1985)
par(mfrow=c(2,1))
hist(wage, freq = FALSE)
hist(log(wage), freq = FALSE)
lines(density(log(wage)), col = 4)

# One categorical variable
summary(occupation)
tab <- table(occupation)
prop.table(tab)
par(mfrow=c(1,2))
barplot(tab)
pie(tab)

# Two categorical variables
xtabs(~ gender + occupation, data = CPS1985)
plot(gender ~ occupation, data = CPS1985)

# Nemerical
tapply(log(wage), gender, mean)
plot(log(wage) ~ education)
plot(log(wage) ~ gender)
detach(CPS1985)