NYT First Said counts, labelling highest

nyt-first-said.R

library(rtweet)
library(tidyverse)
library(glue)

rt <- get_timeline(
  user = "@NYT_first_said",
  n = 18000
)

rt <- rt %>%
  mutate(
    week = round_time(created_at, "weeks"),
    day  = round_time(created_at, "days")
  )

highest_day <- rt %>%
  count(day, sort = TRUE) %>%
  top_n(1) %>%
  pull(day)

bizarre_words <- rt %>% 
  filter(day == highest_day) %>%
  pull(text) %>%
  glue_collapse(sep = ", ")

# there has to be a better way then manually adding line breaks
bw <- structure("bioanthropological, taphonomic, fornici, talentism,oenocentric, \nskateways, seafacing, mchicha, chaza, ufuta, mkate, mishkaki, \nhitokuchigashi, chamoe, vchill, kvavili, kviteli, ojaxuri, gandzili, \nekala, bewk, semispheres, salmiana, sgàilean, retarch, beefcheeks, \nhellasmacked, acousticophobic, epigrammed, demantoids, \nastrocartographer, goosings, hamaka, nonejected, gumtrees, \nreganmian, overincarcerates, vibratode, subfile", class = c("glue", 
                                                                                                                                                                                                                                                                                                                                                                                                                                                                      "character"))
rt_counts <- rt %>%
  group_by(day) %>%
  mutate(
    words = glue_collapse(text, sep = ", "),
    n = n()
    ) %>%
  select(day, words, n) %>%
  distinct(day, .keep_all = TRUE) %>%
  mutate(words = if_else(day == highest_day, bw, words))

ggplot(rt_counts, aes(x = day, y = n)) +
  geom_line(color = "darkgray", size = 0.5) +
  theme_minimal() +
  theme(plot.title = ggplot2::element_text(face = "bold")) +
  labs(
    x = NULL, y = NULL,
    title = "Frequency of tweets from @NYT_first_said",
    subtitle = "Tweet counts aggregated using daily intervals",
    caption = "\nData: Twitter's REST API via rtweet | Graph: @minebocek"    
  ) +
  ggrepel::geom_label_repel(
    data = rt_counts %>% filter(day == highest_day), aes(label = words), 
    hjust = 1.5, vjust = 1
    ) +
  ggrepel::geom_text_repel(
    data = rt_counts %>% filter(day == highest_day), aes(label = as.character(day)),
    vjust = 1
  )