Population pyramid plot with ggplot2 and dplyr (instead of plyr) - r

Question

I am trying to reproduce the simple population pyramid from the post Simpler population pyramid in ggplot2
using ggplot2 and dplyr (instead of plyr).
Here is the original example with plyr and a seed
set.seed(321)
test <- data.frame(v=sample(1:20,1000,replace=T), g=c('M','F'))
require(ggplot2)
require(plyr)
ggplot(data=test,aes(x=as.factor(v),fill=g)) +
geom_bar(subset=.(g=="F")) +
geom_bar(subset=.(g=="M"),aes(y=..count..*(-1))) +
scale_y_continuous(breaks=seq(-40,40,10),labels=abs(seq(-40,40,10))) +
coord_flip()
Works fine.
But how can I generate this same plot with dplyr instead? The example uses plyr in the subset = .(g == statements.
I have tried the following with dplyr::filter but got an error:
require(dplyr)
ggplot(data=test,aes(x=as.factor(v),fill=g)) +
geom_bar(dplyr::filter(test, g=="F")) +
geom_bar(dplyr::filter(test, g=="M"),aes(y=..count..*(-1))) +
scale_y_continuous(breaks=seq(-40,40,10),labels=abs(seq(-40,40,10))) +
coord_flip()
Error in get(x, envir = this, inherits = inh)(this, ...) :
Mapping should be a list of unevaluated mappings created by aes or aes_string

Answer 1

You avoid the error by specifying the argument data in geom_bar:
ggplot(data = test, aes(x = as.factor(v), fill = g)) +
geom_bar(data = dplyr::filter(test, g == "F")) +
geom_bar(data = dplyr::filter(test, g == "M"), aes(y = ..count.. * (-1))) +
scale_y_continuous(breaks = seq(-40, 40, 10), labels = abs(seq(-40, 40, 10))) +
coord_flip()

Answer 2

You can avoid both dplyr and plyr when making population pyramids with recent versions of ggplot2.
If you have counts of the sizes of age-sex groups then use the answer here
If your data is at the individual level (as yours is) then use the following:
set.seed(321)
test <- data.frame(v=sample(1:20,1000,replace=T), g=c('M','F'))
head(test)
# v g
# 1 20 M
# 2 19 F
# 3 5 M
# 4 6 F
# 5 8 M
# 6 7 F
library("ggplot2")
ggplot(data = test, aes(x = as.factor(v), fill = g)) +
geom_bar(data = subset(test, g == "F")) +
geom_bar(data = subset(test, g == "M"),
mapping = aes(y = - ..count.. ),
position = "identity") +
scale_y_continuous(labels = abs) +
coord_flip()

Answer 3

To build an Age Pyramid with individual data or microdata you can use:
test <- data.frame(v=sample(1:100, 1000, replace=T), g=c('M','F'))
ggplot(data = test, aes(x = v, fill = g)) +
geom_histogram(data = subset(test, g == "F"), binwidth = 5, color="white", position = "identity") +
geom_histogram(data = subset(test, g == "M"), binwidth = 5, color="white", position = "identity",
mapping = aes(y = - ..count.. )) +
scale_x_continuous("Age", breaks = c(seq(0, 100, by=5))) +
scale_y_continuous("Population", breaks = seq(-30, 30, 10), labels = abs) +
scale_fill_discrete(name = "Sex") +
coord_flip() +
theme_bw()
Changing the binwidth in geom_histogram() can group your data in wider categories.
Changing binwidth to 10 and adjusting the axis breaks:
ggplot(data = test, aes(x = v, fill = g)) +
geom_histogram(data = subset(test, g == "F"), binwidth = 10, color="white", position = "identity") +
geom_histogram(data = subset(test, g == "M"), binwidth = 10, color="white", position = "identity",
mapping = aes(y = - ..count.. )) +
scale_x_continuous("Age", breaks = c(seq(0, 100, by = 10))) +
scale_y_continuous("Population", breaks = seq(-100, 100, 10), labels = abs) +
scale_fill_discrete(name = "Sex") +
coord_flip() +
theme_bw()