#122
library(Hmisc,T)
library(Design,T)
describe(counties[,-(1:4)])


#124
older <- counties$age6574 + counties$age75
label(older) <- '% age >= 65, 1990'
pdensity <- logb(counties$pop.density+1, 10)
label(pdensity) <- 'log 10 of 1992 pop per 1990 miles^2'

dd <- datadist(counties)
dd <- datadist(dd, older, pdensity)
options(datadist='dd')


#125

v <- varclus(~ pop.density + pop.change + older + crime + college + 
            income + farm + democrat + republican + Perot + 
            white + turnout, data=counties)
plot(v)

win.graph()

s <- summary(democrat ~ pop.density + pop.change + older + crime + 
              college + income + farm + black + white + turnout,
              data=counties)
plot(s, cex.labels=.7)


#127
s <- spearman2(democrat ~ pop.density + pop.change + older + crime + 
              college + income + farm + black + white + turnout,
              data=counties, p=2)
plot(s)


#128
f <- ols(democrat ~ rcs(pdensity,4) + rcs(pop.change,3) + 
  rcs(older,3) + crime + rcs(college,5) + rcs(income,4) + 
  rcs(college,5) %ia% rcs(income,4) + 
  rcs(farm,3) + rcs(white,5) + rcs(turnout,3), data=counties)
f


#130
latex(f)


#131
r <- resid(f)
xYplot(r ~ fitted(f), method='quantile', nx=200, ylim=c(-10,10), xlim=c(20,60), 
	abline=list(h=0, lwd=.5, lty=2), aspect='fill', data=counties)


#132
p1 <-xYplot(r ~ white,    method='quantile', nx=200, ylim=c(-10,10), xlim=c(40,100), 
	abline=list(h=0, lwd=.5, lty=2), aspect = 'fill', data=counties)
p2 <-xYplot(r ~ pdensity, method='quantile', nx=200, ylim=c(-10,10), xlim=c(0,3.5), 
	abline=list(h=0, lwd=.5, lty=2), aspect = 'fill', data=counties)
print (p1, split=c(1,1,1,2), more=T)
print (p2, split=c(1,2,1,2))

qqmath(~r | cut2(pdensity, g=4))


#135
f2<- ols(democrat~(rcs(pdensity,4) + rcs(pop.change,3) + 
  rcs(older,3) + crime + rcs(college,5) + rcs(income,4) + 
  rcs(farm,3) + rcs(white,5) + rcs(turnout,3))^2, data=counties)
f2$stats


#135-136
g<- update(f, x=T)
w <- which.influence(g, 0.3)
dffits <- resid(g, 'dffits')
show.influence(w, data.frame(counties, pdensity, older, dffits), 
	report=c('democrat', 'dffits'), id='county')


#136
an<-anova(f)
an
plot(an, what='partial R2')

#Problem in solvet(cov[idx, idx], coef[idx], tol = tol): apparently singular matrix 
#Use traceback() to see the call stack


#138
incomes <- seq(22900, 32800, length=4)
show.pts <- function(college.pts, income.pt) {
  s <- abs(income - income.pt) < 1650
  x <- college[s]
  x <- sort(x[!is.na(x)])
  n <- length(x)
  low  <- x[10]
  high <- x[n-9]
  college.pts >= low & college.pts <= high
  }
plot(f, college=NA, income=incomes, 
  conf.int=F, xlim=c(0,35), ylim=c(30,55),
  lty=1, lwd=c(.25,1.5,3.5,6), col=c(1,1,2,2),
  perim=show.pts)


#139
plot(f,ylim=c(20,70))

s<-summary(f)
options(digits=4)
plot(s)

f<- Newlabels(f, list(turnout='voter turnout (%)'))


#142
nomogram(f, interact=list(income=incomes), turnout=seq(30,100,by=10),
  lplabel='estiated % voting Democratic',
  ce.var=.8, cex.axis=.75)