Em dash ecology

emdash.R

library(tidyverse)
library(ggtext)
library(openalexR)
library(httr2)


##### OpenAlex #####

query_url <-
  oa_query(entity = "works",
           topics.subfield.id = "subfields/2303",
           has_abstract = TRUE,
           language = "en",
           from_publication_date = "2021-01-01",
           to_publication_date = "2021-12-31",
           type = "article") |>
  paste0("&mailto=XXXXX",
         "&per-page=200",
         "&sample=10000&seed=123")


dat_2021 <- vector("list", 50L)

for (i in 1:(10000/200)) {
  cat("Page:", i)
  query_url_page <- paste0(query_url, "&page=",i)
  req <- request(query_url_page)
  resp <- req_perform(req)
  
  if(resp$status_code != 200) {
    stop("Server error")
  }
  
  json <- resp_body_json(resp)
  
  dat_2021[[i]] <- json$results |>
    works2df() 
}

dat_2021 <- bind_rows(dat_2021)

query_url <-
  oa_query(entity = "works",
           topics.subfield.id = "subfields/2303",
           has_abstract = TRUE,
           language = "en",
           from_publication_date = "2025-01-01",
           to_publication_date = "2025-12-31",
           type = "article") |>
  paste0("&mailto=XXXXX",
         "&per-page=200",
         "&sample=10000&seed=123")


dat_2025 <- vector("list", 50L)

for (i in 1:(10000/200)) {
  cat("Page:", i)
  query_url_page <- paste0(query_url, "&page=",i)
  req <- request(query_url_page)
  resp <- req_perform(req)
  
  if(resp$status_code != 200) {
    stop("Server error")
  }
  
  json <- resp_body_json(resp)
  
  dat_2025[[i]] <- json$results |>
    works2df() 
}

dat_2025 <- bind_rows(dat_2025)


dat <- bind_rows(dat_2021, dat_2025) |> 
  mutate(n_char = nchar(abstract)) |> 
  filter(n_char > 600, n_char < 3500) |> 
  mutate(count_chars = map(abstract, \(x) {
    res <- c(n_em = str_count(x, "—"),
             n_en = str_count(x, "–"),
             n_min = str_count(x, "-"),
             n_aster = str_count(x, "\\*"),
             n_plus = str_count(x, "\\+"),
             n_comma = str_count(x, ","),
             n_col = str_count(x, ":"),
             n_semicol = str_count(x, ";"),
             n_tilde = str_count(x, "~"),
             n_dot = str_count(x, "\\."),
             n_par = str_count(x, "\\("),
             n_question = str_count(x, "\\?"),
             n_slash = str_count(x, "/"),
             n_eq = str_count(x, "="),
             n_perc = str_count(x, "%"),
             n_amp = str_count(x, "&")
    )
    enframe(res, name = "spchar", "n_spchar")
  }, .progress = TRUE)) |> 
  unnest(count_chars)


sp_labs <- c(n_em = "Em dash —",
             n_en = "En dash –",
             n_min = "Minus sign -",
             n_aster = "Asterisk *",
             n_plus = "Plus sign +",
             n_comma = "Comma ,",
             n_col = "Colon :",
             n_semicol = "Semicolon ;",
             n_tilde = "Tilde ~",
             n_dot = "Dot .",
             n_par = "Parenthesis (",
             n_question = "Question mark ?",
             n_slash = "Slash /",
             n_eq = "Equal sign =",
             n_perc = "Percent %",
             n_amp = "Ampersand &") |> 
  enframe(name = "spchar", value = "spchar_lab")


dat |>
  group_by(publication_year, spchar) |> 
  summarise(freq = sum(n_spchar) / sum(n_char)) |> 
  pivot_wider(names_from = publication_year, values_from = freq) |> 
  mutate(ratio = (`2025` - `2021`) /  `2021`) |> 
  left_join(sp_labs) |> 
  ggplot() +
  geom_hline(aes(yintercept = 0L)) +
  geom_col(aes(fct_reorder(spchar_lab, ratio), ratio)) + 
  geom_curve(
    aes(x = x, y = y, xend = xend, yend = yend),
    data = data.frame(x = 12.3, y = 0.75, xend = 16, yend = 0.9),
    curvature = 0.5,
    angle = 90
  ) +
  geom_point(
    aes(x, y), 
    data = data.frame(x = 16, y = 0.9),
    color = "black"
  ) +
  geom_text(
    aes(x, y, label = "Use of Em dash\ndoubled over the period"),
    data = data.frame(x = 12.3, y = 0.75),
    hjust = 0,
    vjust = 1,
    nudge_y = -0.5,
    nudge_x = 0.25,
    size = 3,
  ) +
  coord_flip() +
  scale_y_continuous(limits = c(-1.1, 1.1), labels = scales::percent) +
  labs(title = "The rise of Em dash in ecology article abstracts",
       subtitle = "Change in frequency of use of different characters in the abstracts of ecology articles between 2021 and 2025.",
       caption = "Abstract data: OpenAlex<br>Chart: @francoiskeck.bsky.social") +
  ylab("Usage in 2025 relative to 2021") +
  theme_minimal(base_family = "Lato") +
  theme(axis.title.y = element_blank(),
        plot.title = element_text(
          color = "grey10", 
          size = 16, 
          face = "bold",
          margin = margin(t = 15)
        ),
        plot.subtitle = element_textbox_simple(
          color = "grey30", 
          size = 12,
          lineheight = 1.35,
          margin = margin(t = 15, b = 40)
        ),
        plot.caption = element_markdown(
          lineheight = 1.1,
          color = "grey30", 
          size = 8,
          margin = margin(t = 20)
        ),
        plot.title.position = "plot",
        plot.margin = margin(15, 40, 15, 20),
        plot.background = element_rect(fill = "grey96", color = "grey96"),
        panel.background = element_rect(fill = "grey96", color = "grey96"))