Plotting UCBerkeley admissions using ggplot2. See what the different options in ggplot2 do, and how to get your desired graph.

gistfile1.r

rm(list=ls())
library(datasets)
library(ggplot2)
str(UCBAdmissions) #Table, but ggplot needs data frames


ucb <- data.frame(UCBAdmissions)
names(ucb)
head(ucb)

#We Would Like to Create some Barplots
p <- ggplot(ucb, aes(x=Gender, y=Freq) ) 
histg<- geom_bar(stat ="identity")
p+histg # All the departments and Admit/Rejects are mushed together

#We need to separate it out

#First, pull the departments apart
p + histg + facet_grid(. ~ Dept) #Accept and Reject still need to be separated out

p + histg + facet_grid(. ~ Dept+Admit) # would be nice to color them differently. Use fill for that

#Let's use fill
p2 <- ggplot(ucb, aes(x=Gender, y=Freq, fill=factor(Admit)) ) 

p2 + histg + facet_grid(. ~ Dept + Admit) #not good. Because it is difficult to compare same gender

#let's try
p2 + histg + facet_grid(. ~ Admit+Dept) #This made it worse. The departments are too far apart.

#How about
p2 + histg + facet_grid(. ~ Dept)  #Almost there. But we don't want a stacked barplot

#But we want Male Admits and Rejects together for each department.

#Notice that X aesthtic, the gender is the closest together.

#This means that in our final plot, ADMIT will have to be the x-aesthetic
#start again
p3 <- ggplot(ucb, aes(x=Admit, y=Freq, fill=factor(Admit)) ) 

p3 + histg + facet_grid(. ~ Dept+Gender)  

p3 + histg + facet_grid(. ~ Dept+Gender)  + theme(axis.text.x=element_blank())

#Finally!