CHAMP literature review figure 3: mean of medians curves

attach(girlsA)
a<-aggregate(girlsA$sums,list(girlsA$age),sum)
b<-aggregate(girlsA$n,list(girlsA$age),sum)
c<-merge(a,b,by="Group.1")
c$BMI<-c$x.x/c$x.y
c$gender<-"High tracking"
dataB<-data[data$group=="B",]
dataB$wts<-1/dataB$n
attach(girlsB)
d<-aggregate(girlsB$sums,list(girlsB$age),sum)
e<-aggregate(girlsB$n,list(girlsB$age),sum)
f<-merge(d,e,by="Group.1")
f$BMI<-f$x.x/f$x.y
f$gender<-"Low tracking"
dataC<-data[data$group=="C",]
dataC$wts<-1/dataC$n
attach(girlsC)
g<-aggregate(girlsC$sums,list(girlsC$age),sum)
h<-aggregate(girlsC$n,list(girlsC$age),sum)
i<-merge(g,h,by="Group.1")
i$BMI<-i$x.x/i$x.y
i$gender<-"Outlying"
dataD<-data[data$group=="D",]
dataD$wts<-1/dataD$n
attach(girlsD)
j<-aggregate(girlsD$sums,list(girlsD$age),sum)
k<-aggregate(girlsD$n,list(girlsD$age),sum)
l<-merge(j,k,by="Group.1")
l$BMI<-l$x.x/l$x.y
l$gender<-"WHO"
dataE<-data[data$group=="E",]
dataE$wts<-1/dataE$n
attach(girlsE)
m<-aggregate(girlsE$sums,list(girlsE$age),sum)
n<-aggregate(girlsE$n,list(girlsE$age),sum)
o<-merge(m,n,by="Group.1")
o$BMI<-o$x.x/o$x.y
o$gender<-"CDC"
df<-rbind(c,f,i,l,o)
df$std<-ifelse(df$gender=="CDC"|df$gender=="WHO",paste("std"),paste("not"))
df$age<-df$Group.1
attach(df)
library(ggplot2)
PL<-ggplot(df, aes(x=age, y=BMI, xlim(2,11),colour=gender, shape=gender,linetype=std)) +geom_smooth(se=TRUE)+
  labs(x="Age (yrs)",y="BMI")+theme_bw()+scale_x_continuous(breaks = seq(1, 28, 0.5))+
  scale_y_continuous(breaks = seq(14, 30, 0.4))+ theme(legend.title=element_blank())
cbbPalette<-c("Black","Red","Blue","grey84","azure4")
PL+scale_colour_manual(values=cbbPalette)