zc = function() {
    y3 = y[,3]			# zipcodes
    y3[is.na(y3)] = y[1,3]
    kk = unique(y3)
    n = length(y3)
    nk = length(kk)
    ii = kk
    for(i in 1:nk) { ii[i] = seq(n)[y3==kk[i]][1] }
      	 ## now y[ii,] contains 1st instance of a zip code record
}


zc = read.table("zipcodes.txt",TRUE)	# 394 unique zip codes + long/lat
zcp = rep(0,100000)
for(i in 1:394) { zcp[zc[i,1]] = i }	# pointer from zipcode to zc row

x.zc = x[,3];  iomit = seq(n)[is.na(x.zc)]; nn = length(x.zc)
x.zc = x[-iomit,3]
x.pr = log10(x[-iomit,5])
x.dt = x[-iomit,10]

       # price by zip code
pr.by.zc = rep(0,394); for(i in 1:394) { pr.by.zc[i] = mean( x.pr[ x.zc == zc[i,1] ] ) }
p2.by.zc = rep(0,394); for(i in 1:394) { p2.by.zc[i] = median( x.pr[ x.zc == zc[i,1] ] ) }
	 plot(p2.by.zc,pr.by.zc,pch=16)
   cex.pr = 0.5 + 1.5 * ( ( pr.by.zc - min(pr.by.zc,na.rm=TRUE) ) / diff(range(pr.by.zc,na.rm=TRUE)) )
     plot( zc[,2],zc[,3], pch=16, cex=cex.pr )
   	 library("maps"); map('county','california',add=T)
     plot( zc[,2],zc[,3], pch=16, cex=cex.pr,xlim=c(-122.7,-122.2),ylim=c(37.5,38.0) )
   col.pr = trunc( 1 + 31 * ( ( pr.by.zc - min(pr.by.zc,na.rm=TRUE) ) / diff(range(pr.by.zc,na.rm=TRUE)) ) )
     plot( zc[,2],zc[,3], pch=16, col=rainbow(32)[col.pr] )
         map('county','california',add=T)
     plot( zc[,2],zc[,3], pch=16, col=rainbow(32)[col.pr],xlim=c(-122.7,-122.2),ylim=c(37.5,38.0) ) 



	# price by date

x.yr = as.integer(substr(x.dt,1,4))
x.mo = as.integer(substr(x.dt,6,7))

    hist(x.yr,seq(2002.5,2008.5))
    hist(x.mo,seq(.5,12.5))

pr.yr = rep(0,6)
for(i in 2003:2008) {pr.yr[i-2002] = mean( x.pr[x.yr==i] )}
      plot(2003:2008,pr.yr,type="b",pch=16)
      text(c(2003.5,2006),log10(c(465000,661000)),c("$465,000","$661,000"))

pr.mo = rep(NA,72); k=0; yr.mo = seq( 2003+1/24,2009,1/12 )
for(i in 2003:2008) { for( j in 1:12 ) { k = k+1; m = seq(nn)[x.yr==i & x.mo==j]; pr.mo[k]=mean(x.pr[m],na.rm=TRUE) } }
      plot(yr.mo,pr.mo,type="b",pch=16)
      text(c(2008,2006.8),log10(c(377000,697000)),c("$377,000","$697,000"))



	# price by date  and  zip code

zc.list = c(94957,94528,93920,94513,94565,94533,94509)
      plot(yr.mo,pr.mo,type="b",pch=16,ylim=c(5,6.5))
      text(c(2008,2006.8),log10(c(377000,697000)),c("$377,000","$697,000"))
for(iz in 1:length(zc.list)) { z = zc.list[iz]
    pr.mo.z = rep(NA,72); k=0
    for(i in 2003:2008) { for( j in 1:12 ) { k = k+1; m = seq(nn)[x.yr==i & x.mo==j & x.zc==z]
        pr.mo.z[k]=mean(x.pr[m],na.rm=TRUE) } }
    lines(yr.mo,pr.mo.z); text( 2002.5+iz,5.2,as.character(z) )
}