library(babynames)
library(reshape2)
library(ineq)

# babynames is based on SSA applicant number and sex.
# we estimate that the sex-ratio at birth is 107 boys for 100 girls
# we correct the babynames data

applicants<-dcast(applicants,year~sex,sum)
applicants$CORR<-107/(100*applicants$M/applicants$F)

bebe<-dcast(babynames,name+year~sex,value.var="n",sum)
m<-match(bebe$year,applicants$year)
bebe$FCORR<-round(bebe$F/applicants$CORR[m])

bebe$PROP<-bebe$M/(bebe$FCORR+bebe$M) # 1 if every baby is male, 0 if female

bebe$PROPM<-bebe$M/applicants$M[m]
bebe$PROPF<-bebe$FCORR/(applicants$F[m]/applicants$CORR[m])



# test sur Leslie
pr<-"Leslie"
	test <- subset(bebe,bebe$name==pr)
    test$PROP<-test$M/test$FCORR	
	par(mar=c(4, 6, 2, .2) + 0.1)
    plot(test$year,log10(test$PROP),main=pr,pch=20,cex=.5,ylim=c(-2,2),axes=F,ylab="",xlab="")
    
    smoothingSpline = smooth.spline(test$year,log10(test$PROP), spar=0.4)
    lines(smoothingSpline)
    axis(1)
axis(2, at=c(-2,-1,0,1,2),col.axis="black", las=2,labels=c("100 times \n more girls","10 times\n more girls","equal number","10 times\nmore boys","100 times\n more boys"))
title(sub="Data : USA, Social Security Administration, package babynames | coulmont.com",cex.sub=.8,adj=0,line=3)


# more epicene

mini<-.1
maxi<-.9
bbe<-subset(bebe,bebe$PROP>mini&bebe$PROP<maxi)

bbe$TOTAL<-50*(bbe$PROPF+bbe$PROPM) # sum(bbe$F) is around 1 and sum(bbe$M) is around 1 too
bbp<-aggregate(TOTAL~year,data=bbe,sum)
par(mar=c(5,4,4,4))
plot(bbp$year,bbp$TOTAL,axes=F,pch=20,cex=.85,ylim=c(0,8),ylab="% babies with androgynous forename",xlab="",col="lightblue")
axis(2, las=2)
axis(1)
smoothingSpline = smooth.spline(bbp$year,bbp$TOTAL, spar=0.4)
lines(smoothingSpline,lwd=2)
title(main="An increasing proportion of babies have androgynous names\n...and the number of androgynous names is also increasing")
title(sub="Data : USA, Social Security Administration, package babynames | coulmont.com\nLeft scale : Proportion | Right scale : Number of androgynous names\nNote: Androgynous name = name for which female are more than 10% and less than 90% of the total",cex.sub=.8,adj=0)

bbe$NB<-1
bbn<-aggregate(NB~year,data=bbe,sum)
smoothingSpline = smooth.spline(bbn$year,.005*bbn$NB, spar=0.4)
lines(smoothingSpline,lwd=1,lty=3)
#lines(bbn$year,.01*bbn$NB,lty=3)
axis(4,las=2,at=c(0,2.5,5,7.5),labels=c(0,500,1000,1500))



## GINI

bbe$NB<-bbe$M+bbe$FCORR
bbe$N<-1

bbm<-aggregate(M~year,data=bbe,sum)
bbm$F<-aggregate(FCORR~year,data=bbe,sum)$FCORR

bbm$PROPM<-100*bbm$M/(bbm$F+bbm$M)



gini<-NULL
for (i in 1880:2013) {
	gini[i-1879]<-ineq(subset(bbe,bbe$year==i)$NB,type="Gini")
}
gini<-as.data.frame(gini)
gini$year<-1880:2013

# proportion of androgynous names that are male names (given mostly to boys)
bbe$PRM<-as.numeric(bbe$M>bbe$FCORR) # 1 if this name is given to boys more than to girls
bbnb<-aggregate(N~year,data=bbe,sum)
bbnb$PRM<-aggregate(PRM~year,data=bbe,sum)$PRM
bbnb$PRPRM<-100*bbnb$PRM/bbnb$N

plot(bbm$year,bbm$PROPM,col="orange",pch=20,cex=.5,ylim=c(0,100),yaxt="n",ylab="Proportion",xlab="",axes=F)
axis(2,las=2,at=c(0,25,50,75,100),labels=c(0,25,50,75,100))
axis(1)
smoothingSpline = smooth.spline(bbm$year,bbm$PROPM, spar=0.4)
lines(smoothingSpline)
abline(50,0,lty=3,col="darkgray")
smoothingSpline = smooth.spline(bbnb$year,bbnb$PRPRM, spar=0.4)
lines(smoothingSpline,col="red",lty=2)

title(main="Sex ratio")
title(sub="Data : USA, Social Security Administration, package babynames | coulmont.com\nBetween 1950 and 1990, boys received more girls' names than girls boys' names\nNote: Androgynous name = name for which female are more than 10% and less than 90% of the total",cex.sub=.8,adj=0)
text(1890,65,"1900: 60% of epicene names\nare male names",cex=.8,font=2,col="red",adj=0)
text(1970,32,"1980: 35% of epicene babies\nare boys",cex=.8,font=2,col="black",adj=0)



par(new=TRUE)
plot(gini$year,gini$gini,type="n",xaxt="n",yaxt="n",xlab="",ylab="",axes=F)
smoothingSpline = smooth.spline(gini$year,gini$gini, spar=0.4)
lines(smoothingSpline,col="violet",lty=2,lwd=2)
axis(4,las=2,col="violet")
mtext("Gini coefficient",side=4,line=3,col="violet",font=2)