Are movies that win the Oscars really longer than in the past? Nah... (slow at first run, but data are cached locally)

oscar_winners_length.R

# https://en.wikipedia.org/wiki/List_of_Academy_Award-winning_films

library("dplyr", warn.conflicts = FALSE)
library("tidywikidatar")
library("ggplot2")
# https://github.com/ivelasq/severance
library(severance)

tw_set_language(language = "en") 

tw_enable_cache()
tw_create_cache_folder()
tw_set_cache_folder(path = fs::path(fs::path_home_r(),
                                    "R",
                                    "tw_data_oscar"))

linked_df <- tw_get_wikipedia_page_links(url = "https://en.wikipedia.org/wiki/List_of_Academy_Award-winning_films")

duration_df <- linked_df %>% 
  dplyr::select("qid") %>% 
  dplyr::distinct(qid, .keep_all = TRUE) %>% 
  dplyr::mutate(title = tw_get_label(id = qid),
                date = tw_get_p1(id = qid, p = "P577"),
                duration = tw_get_p1(id = qid, p = "P2047")) %>% 
  dplyr::filter(is.na(date)==FALSE) %>% 
  mutate(date = stringr::str_extract(string = date, pattern = "[[:digit:]]+") %>% as.numeric(),
         duration = stringr::str_extract(string = duration, pattern = "[[:digit:]]+") %>% as.numeric()) %>% 
  dplyr::filter(duration>60)


duration_df  %>% 
  #dplyr::filter(date>1945) %>% 
  ggplot(mapping = aes(x = date, y = duration)) +
  geom_point() +
  geom_smooth(method = "lm")

duration_df  %>% 
  dplyr::mutate(decade = stringr::str_extract(string = date, pattern = "[[:digit:]][[:digit:]][[:digit:]]")) %>% 
  dplyr::group_by(decade) %>% 
  dplyr::summarise(median_by_decade = median(duration)) %>% 
  dplyr::ungroup() %>% 
  dplyr::mutate(decade = stringr::str_c(decade, "0s")) %>% 
  dplyr::arrange(dplyr::desc(median_by_decade)) %>% 
  #dplyr::filter(date>1945) %>% 
  ggplot(mapping = aes(x = median_by_decade,
                       y = decade,
                       label = scales::number(median_by_decade, accuracy = 1),
                       fill = decade)) +
  geom_col()  +
  geom_text(hjust = 1.2, colour = "white", family = "Roboto Condensed") +
  scale_y_discrete("") +
  scale_x_continuous(name = "") +
  scale_fill_manual(values = rep(x = severance_palette("Jazz02"),2)) +
  #scale_colour_identity() +
  theme_minimal(base_family = "Roboto Condensed",
                base_size = 16) +
  theme(legend.position = "none") +
  labs(title = "Median running time of Academy Award-winning films",
       subtitle = "Only films longer than 60 minutes included.\nDuration expressed in minutes. Colour of bars by decade.",
       caption = "Source: https://en.wikipedia.org/wiki/List_of_Academy_Award-winning_films")

ggsave(filename = "duration_of_films_barchart.png",
       width = 7, height = 7,bg = "white")



duration_df  %>% 
  dplyr::mutate(decade = stringr::str_extract(string = date, pattern = "[[:digit:]][[:digit:]][[:digit:]]")) %>% 
  dplyr::group_by(decade) %>% 
  dplyr::mutate(median_by_decade = median(duration)) %>% 
  dplyr::ungroup() %>% 
  dplyr::mutate(decade = stringr::str_c(decade, "0s")) %>% 
  dplyr::arrange(dplyr::desc(median_by_decade)) %>% 
  dplyr::filter(duration<300) %>% 
  #dplyr::filter(date>1945) %>% 
  ggplot(mapping = aes(x = duration,
                       y = date,
                       label = scales::number(median_by_decade, accuracy = 1),
                       colour = decade)) +
  geom_point()  +
  #geom_text(hjust = 1.2, colour = "white", family = "Roboto Condensed") +
  scale_y_continuous("", n.breaks = 10) +
  scale_x_continuous(name = "") +
  scale_colour_manual(values = rep(x = severance_palette("Jazz02"),2)) +
  #scale_colour_identity() +
  theme_minimal(base_family = "Roboto Condensed",
                base_size = 16) +
  theme(legend.position = "none") +
  labs(title = "Median running time of Academy Award-winning films",
       subtitle = "Only films longer than 60 and shorter than 300 minutes included.\nDuration expressed in minutes. Colour of dots by decade.",
       caption = "Source: https://en.wikipedia.org/wiki/List_of_Academy_Award-winning_films")

ggsave(filename = "duration_of_films_scatter.png",
       width = 7, height = 7,bg = "white")

Resulting charts: