Charting Faculty Salaries

analysis.r

library(dplyr)
library(reshape2)
library(stringr)
library(ggplot2)
library(RColorBrewer)

# options(stringsAsFactors = FALSE)
options("scipen"=100, "digits"=4)

salaries <- read.csv("faculty-salaries-2013-14.csv")

salaries <- salaries %.%
  melt(id = c("rank", "discipline"),
       variable.name = "institution_type",  value.name = "salary")

avg_salaries <- salaries %.%
  filter(discipline == "All disciplines combined")

salaries$discipline <- salaries$discipline %.% str_sub(1,32)

full_salary <- salaries %.%
  filter(rank == "Professor") %.%
  filter(institution_type == "All_institutions") %.%
  arrange(salary)

salary_labels <- filter(salaries, rank == "Professor" | 
                          rank == "Assistant" |
                          rank == "Associate")

salaries$discipline <- ordered(salaries$discipline,
                               levels = full_salary$discipline)

salaries$rank <- ordered(salaries$rank,
                         levels = c("New assistant",
                                    "Assistant",
                                    "Associate",
                                    "Professor"))

breaks <- seq(0e3, 165e3, 10e3)
labels <- breaks %.% str_replace("(.+)000", "\\1K")

nice_thousands <- function(value) {
  text <- round(value / 1e3, digits = 0) * 1000
  text <- str_replace(as.character(text), "(.+)000", "\\1K")
  return(text)
}

attribution <- function(footnoteText = format(Sys.time(), "%d %b %Y"),
                        size= 0.9, 
                        color= "gray8") {
  require(grid)
  pushViewport(viewport())
  grid.text(label= footnoteText ,
            x = unit(1,"npc") - unit(2, "mm"),
            y= unit(2, "mm"),
            just=c("right", "bottom"),
            gp=gpar(cex= size, col=color))
  popViewport()
}

plot_institution <- ggplot(data = salaries) +
  geom_bar(aes(x = discipline, y = salary, fill = rank, group = NA),
           stat = "identity", position = "dodge") +
  facet_wrap(~ institution_type, ncol = 1) +
  coord_flip() +
  theme(legend.position = "bottom",
        axis.text.y = element_text(color = "black", size = rel(1.1)),
        strip.text = element_text(color = "black", size = rel(1.25)),
        title = element_text(size = rel(1.25))) +
  scale_y_continuous(breaks = breaks, labels = labels) +
  geom_hline(yintercept = breaks, color = "white") + 
  geom_hline(data = avg_salaries, aes(yintercept = salary),
             size = 1, color = "gray18") +
  geom_text(data = salary_labels,
            aes(x = discipline, y = (salary - 2.1e3), fill = rank,
                label = nice_thousands(salary)),
            color = "gray8", size = 2.5) +
  xlab(NULL) + ylab(NULL) +
  scale_fill_brewer(palette = "Paired") +
  ggtitle("Faculty Salaries by Discipline, Rank, and Institution Type, 2013-14")

plot_discipline <- ggplot(data = salaries) +
  geom_bar(aes(x = institution_type, y = salary, fill = rank, group = NA),
           stat = "identity", position = "dodge") +
  facet_wrap(~ discipline, ncol = 1) +
  coord_flip() +
  theme(legend.position = "bottom",
        axis.text.y = element_text(color = "black", size = rel(1.1)),
        strip.text = element_text(color = "black", size = rel(1.25)),
        title = element_text(size = rel(1.5))) +
  scale_y_continuous(breaks = breaks, labels = labels) +
  geom_hline(yintercept = breaks, color = "white") + 
  geom_text(data = salary_labels,
            aes(x = institution_type, y = (salary - 2.1e3), fill = rank,
                label = nice_thousands(salary)),
            color = "gray8", size = 2.5) +
  xlab(NULL) + ylab(NULL) +
  scale_fill_brewer(palette = "Paired") +
  ggtitle("Faculty Salaries by Discipline, Rank, and Institution Type, 2013-14")

svg("faculty-salaries-by-discipine.svg", width = 14, height = 32)
  print(plot_discipline)
  attribution("Data: Chronicle of Higher Education; Visualization: Lincoln Mullen <http://lincolnmullen.com>")
dev.off()

svg("faculty-salaries-by-institution.svg", width = 14, height = 38)
  print(plot_institution)
  attribution("Data: Chronicle of Higher Education; Visualization: Lincoln Mullen <http://lincolnmullen.com>")
dev.off()

pdf("faculty-salaries-by-institution.pdf", width = 14, height = 32)
  print(plot_institution)
  attribution("Data: Chronicle of Higher Education; Visualization: Lincoln Mullen <http://lincolnmullen.com>")
dev.off()

pdf("faculty-salaries-by-discipline.pdf", width = 14, height = 38)
  print(plot_discipline)
  attribution("Data: Chronicle of Higher Education; Visualization: Lincoln Mullen <http://lincolnmullen.com>")
dev.off()

faculty-salaries-2013-14.csv