econ_plot.R

library(tidyverse)

# read data
dat_raw <- read_csv('http://infographics.economist.com/databank/Economist_women-research.csv')

# drop stuff that's not the data
dat <- dat_raw[2:13,]

# fix names
names(dat) <-
  dat_raw[1,] %>%
  t %>%
  unname %>%
  tolower %>%
  sub('%.*', 'inventors', .) %>%
  gsub('science(s|\\b)?', 'sci', .)

# convert to long format. melt() forever, by the way -- sorry, tidyverse.
dat <- data.table::melt(dat, id.var='country')
dat$value <- as.numeric(dat$value)

# a bunch of mutations
dat <-
  dat %>% mutate(
    # group countries
    country2 = 
      ifelse(country %in% c('Portugal', 'Japan', 'Brazil'), 
             country, 'Others*') %>%
      factor(levels = rev(c('Japan', 'Others*', 'Brazil', 'Portugal')),
             ordered = T),
    # field of study to factor
    variable = as.factor(variable)) %>%
  
  # add offset for same-value collisions
  group_by(variable, value) %>%
  mutate(offset = dense_rank(country) - 1)

# economist-type colors, courtesy ggthemes
get_econ_clr <- function(clr_name) {
  filter(ggthemes::ggthemes_data[["economist"]]$fg, 
         name == clr_name)$value[1]
}
plot_clrs <- c(Japan =     get_econ_clr('pink'),
               Portugal =  get_econ_clr('dark blue'),
               Brazil =    get_econ_clr('light blue'),
               `Others*` = get_econ_clr('gray'))

# make the plot. all the hard coded size values are cherry-picked/eyeballed 
# for a specific plot window size

dat %>%
  # drop this category
  filter(variable != 'inventors') %>%
  
  # this weird aes(x) call is so dots will stack up
  ggplot(aes(x=as.numeric(variable) + offset/7, y=value)) +
  
  # points and an 50% reference line
  geom_point(size=1.5, aes(color=country2)) +
  geom_hline(yintercept = .5) +
  
  # make it wide not tall
  coord_flip() +
  
  scale_color_manual(values = plot_clrs) +
  scale_x_continuous(breaks = seq_along(levels(dat$variable)),
                     labels = levels(dat$variable),
                     expand = c(.05,.05)) +
  scale_y_continuous(expand = c(0, 0), limits = c(0,1),
                     labels = scales::percent) +
  theme_minimal() +
  theme(plot.title = element_text(hjust=3.5),
        panel.grid.minor.y = element_blank(),
        plot.caption = element_text(color='grey50')) +
  labs(title = "Share of published researchers who are women, by field & country",
       caption = paste0("*", other_countries),
       x=NULL, y=NULL, color=NULL)

Result: