babynames.R

a=read.table('usa_big.txt', header=F)
remove_and_make_numeric=function(s){
  return(as.numeric(substr(x = s, start = 1, stop = nchar(s)-1)))
}

summarize=function(x){
  y=x/100
  within=sum(y)
  res=0
  for(i in 1:length(y)){
    res=res+y[i]^2
  }
  res=res#+within^2/(500-20)
  return(res)
}

calc_same_name_prob=function(xnames, xprobs, ynames, yprobs){
  y_indexes_to_keep=which(ynames!='.')
  x_indexes_to_keep=which(xnames!='.')
  xprobs=xprobs[x_indexes_to_keep]
  xnames=xnames[x_indexes_to_keep]
  yprobs=yprobs[y_indexes_to_keep]
  ynames=ynames[y_indexes_to_keep]
  xlist=as.list(xprobs)
  names(xlist) <- xnames
  #print(xlist)
  res=0
  contribs=c()
  names_c=c()
  for(i in 1:length(ynames)){
    if(ynames[i] %in% xnames){
      #print(xlist[ynames[i]])
      res=res+yprobs[i]*as.numeric(xlist[ynames[i]])/10000
      names_c=c(names_c, ynames[i])
      contribs=c(contribs,yprobs[i]*as.numeric(xlist[ynames[i]])/10000)
    }
  }
  if(length(names_c)>0){
    index=which.max(contribs)
    print(paste(names_c[index],';',contribs[index]))
  }
  return(res)
}


colnames(a) <- c('rank', 'boy', 'boyperc', 'girl', 'girlperc')
a$boyperc <- sapply(as.character(a$boyperc), remove_and_make_numeric)
a$girlperc <- sapply(as.character(a$girlperc), remove_and_make_numeric)
a$year=rep(2016:1880, each=1000)

same_name=rep(0,nrow(a)/1000)
for(i in 1:(nrow(a)/1000)){
  is=((i-1)*1000+1):(i*1000)
  #print(is)
  same_name[i]=calc_same_name_prob(as.character(a$boy)[is], a$boyperc[is],
                                   as.character(a$girl)[is], a$girlperc[is])
  #cat('.')
}

b=reshape(a,  drop=c('girl', 'boy'),idvar=c('year'),v.names=c('boyperc','girlperc'),timevar='rank', direction='wide')
b$top10_boys=apply(b[,c(1:1000)*2],1,summarize)
b$top10_girls=apply(b[,c(1:1000)*2+1],1,summarize)
plot(b$year, b$top10_boys, col='blue', type='l')
points(b$year, b$top10_girls, col='red', type='l')
essential_1=b[,c(1,ncol(b)-1)]
essential_2=b[,c(1,ncol(b))]
new_df=as.data.frame(rbind(as.matrix(essential_1), as.matrix(essential_2)))
colnames(new_df) <- c('Jaro','Probableco')
new_df$Sekso=rep(c('Knabo','Knabino'), each=nrow(new_df)/2)
library(ggplot2)

xticks=c(1880+0:6*20,2016)
ggplot(data = new_df, aes(x=Jaro, y=Probableco, color=Sekso))+
  geom_line()+ggtitle('Probableco, ke du beboj havas identajn personajn nomojn en Usono')+
  scale_color_manual(values=c('red',"blue"))+ylim(c(0,0.025))+
  scale_x_continuous(name= 'Naskigxjaro', breaks=xticks, limits=c(1880,2016))+theme_bw()