Trying to recreate 538 Bechdel Test plot

bechdel_538

# plot is here: https://fivethirtyeight.com/features/the-dollar-and-cents-case-against-hollywoods-exclusion-of-women/

library(tidyverse)
library(fivethirtyeight)
library(lubridate)
library(scales)

extract_period <- function(year) {
  period <- year - year %% 5
  century <- year - year %% 100
  period_right <- period + 4 - century
  period_right <- ifelse(period_right < 10, str_c("0", period_right), period_right)
  str_c(period, "-\n'", period_right)
}

# non-working version
bechdel_summarized <- bechdel %>%
  mutate(period = extract_period(year),
         binary2 = case_when(
           clean_test %in% c("ok", "dubious") ~ "PASS",
           TRUE ~ "FAIL"
         )) %>%
  count(period, clean_test, binary2) %>%
  group_by(period, binary2) %>%
  summarise(nn = sum(n)) %>%
  group_by(period) %>%
  mutate(pct = nn/sum(nn)) %>%
  filter(binary2 == "PASS") %>%
  select(period, pct)

bechdel %>%
  mutate(period = extract_period(year)) %>%
  count(period, clean_test, binary) %>%
  inner_join(bechdel_summarized, by = "period") %>%
  ggplot(aes(x = period, fill = clean_test)) +
  geom_bar(aes(y = n), position="fill", stat="identity", color = "#f0f0f0", width = 0.98, size=1) +
  geom_step(aes(y = pct, group = 1), lwd = 2) +
  labs(title = "The Bechdel Test Over Time",
       subtitle = "How women are represented in movies",
       caption = "SOURCE: BECHDELTEST.COM",
       y = NULL, x = NULL) +
  guides(fill = FALSE) +
  scale_y_continuous(labels = c("0", "25", "50", "75", "100%"),
                     expand = c(0, 0),
                     sec.axis = sec_axis(~ ., breaks = c(0.25, 0.5, 0.6, 0.8, 0.96),
                                         labels = c("Passes\nBechdel\nTest", "Dubious", "Women only\ntalk about men", "Women don't\ntalk to each\nother", "Fewer than\ntwo women"))) +
  theme_classic() +
  scale_x_discrete(
    breaks = c("1970-\n'74", "1980-\n'84", "1990-\n'94", "2000-\n'04", "2010-\n'14"),
    expand = c(0, 0)) +
  scale_fill_manual(values=c("#ff2600", "#ff937f", "#ffcac0", "#6ab2d5", "#008fd5")) +
  theme(axis.ticks.x = element_blank(),
        plot.background = element_rect(fill = "#f0f0f0"),
        panel.grid = element_blank(),
        panel.border = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks.length=unit(0.3, "cm")) +
  annotate("text", x = 5, y = 0.75, label= "bold(FAIL)", parse = TRUE, size = 12) +
  annotate("text", x = 5, y = 0.2, label= "bold(PASS)", parse = TRUE, size = 12)

### working version although not "nice", after getting help from Twitter

bechdel_by_time <- bechdel %>%
  mutate(period = extract_period(year)) %>%
  count(period, clean_test, binary) %>%
  inner_join(bechdel_summarized, by = "period")

# need to have an additional DF where the last percent is added with an additional period level
# for the geom_step to work
last_pct <- bechdel_by_time$pct[length(bechdel_by_time$pct)]
bechdel_add_step <- tibble(period = c(bechdel_by_time$period, "another_level"), pct = c(bechdel_by_time$pct, last_pct))

ggplot(data = NULL) +
  geom_bar(data = bechdel_by_time, aes(x = period, y = n, fill = clean_test), position="fill", stat="identity", color = "#f0f0f0", width = 0.98, size=1) +
  geom_step(data = bechdel_add_step, aes(x = period, y = pct, group = 1), lwd = 2, position = position_nudge(-0.5)) +
  labs(title = "The Bechdel Test Over Time",
       subtitle = "How women are represented in movies",
       caption = "SOURCE: BECHDELTEST.COM",
       y = NULL, x = NULL) +
  guides(fill = FALSE) +
  scale_y_continuous(labels = c("0", "25", "50", "75", "100%"),
                     expand = c(0, 0),
                     sec.axis = sec_axis(~ ., breaks = c(0.25, 0.5, 0.6, 0.8, 0.96),
                                         labels = c("Passes\nBechdel\nTest", "Dubious", "Women only\ntalk about men", "Women don't\ntalk to each\nother", "Fewer than\ntwo women"))) +
  theme_classic() +
  scale_x_discrete(
    breaks = c("1970-\n'74", "1980-\n'84", "1990-\n'94", "2000-\n'04", "2010-\n'14"),
    expand = c(0, -1)) +
  scale_fill_manual(values=c("#ff2600", "#ff937f", "#ffcac0", "#6ab2d5", "#008fd5")) +
  theme(axis.ticks.x = element_blank(),
        plot.background = element_rect(fill = "#f0f0f0"),
        panel.grid = element_blank(),
        panel.border = element_blank(),
        axis.line.y = element_blank(),
        axis.ticks.length=unit(0.3, "cm"),
        plot.caption = element_text(size = 8, hjust = 1.3)) +
  annotate("text", x = 5, y = 0.75, label= "bold(FAIL)", parse = TRUE, size = 12) +
  annotate("text", x = 5, y = 0.2, label= "bold(PASS)", parse = TRUE, size = 12)