Gender parity in HarvardX and MITx courses (2012-13)

genderparity.R

# data: http://thedata.harvard.edu/dvn/dv/mxhx
# notes: country is determined by mix of administrative (computed from IP address) and user-provided (filled in from student address if available when IP was indeterminate); during de-identification, some country names were replaced with the corresponding continent/region name.

setwd("~/Desktop/edX")

library('lattice')
library('ggplot2')

data <- read.csv('HMXPC13_DI_v2_5-14-14.csv')

d <- data

# get count by female
f <- d[grepl('f', d$gender), ]
female <- as.data.frame(table(f$final_cc_cname_DI))
names(female) <- c('country', 'female')

# get count by male
# TODO: this is SILLY there has to be a better way of aggregating by gender
m <- d[grepl('m', d$gender), ]
male <- as.data.frame(table(m$final_cc_cname_DI))
names(male) <- c('country', 'male')

# merge male and female counts by country column
total <- merge(female,male,by='country')
total <- total[total$male > 0, ]
total$ratio <- total$female/total$male

# plot
ggplot(total, aes(x = reorder(country,ratio),)) +
  geom_point(stat="identity", ymin=0, aes(y=ratio, ymax=ratio)) +
  geom_text(aes(x=country,
                y=ratio,
                ymin=-0.05,
                ymax=ratio,
                label=country,
                hjust=1)) +
  theme(axis.text.y = element_blank(),
        axis.ticks.y = element_blank(),
        axis.title.y = element_blank()) +
  ylab('gender parity') +
  coord_flip() +
  ggtitle('Gender parity in HarvardX and MITx courses (2012-13)')