jobmarket_plots.R

library(tidyverse)
library(lubridate)
d <- read_csv("psychjobwiki.csv")

ms <- d %>%
  group_by(year, category) %>%
  count() %>%
  left_join(d %>% group_by(year) %>% count() %>% rename(total = n)) %>%
  mutate(prop = n / total)

ggplot(ms, aes(x = year, y = prop, col = category)) + 
  geom_point(aes(size = n), alpha = .3) + 
  geom_smooth(se = FALSE, method = "lm") + 
  ggthemes::theme_few() + 
  theme(legend.position = "bottom") + 
  xlab("Year") + 
  ylab("Proportion of Total Jobs") + 
  scale_size_area(guide = FALSE) + 
  ggthemes::scale_color_ptol(name = "")

d$deadline <- as_date(d$deadline)
d$months <- month(d$deadline)

ggplot(d, aes(x = months, fill = year)) + 
  geom_histogram(binwidth = 1, aes(y = ..density..)) + 
  ggthemes::theme_few() + 
  facet_wrap(~ year)+ 
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = .5)) + 
  geom_vline(data = d %>% group_by(year) %>% summarise(months = mean(months)), 
                                                       aes(xintercept = months), 
             col = "red", lty = 2) + 
  scale_x_continuous(breaks = c(0,3,6,9,12), name = "Deadline Month")