code for an `rtweet` analysis of tweets from SACNAS 2019

sacnas-rtweet.R

library(rtweet)
library(tidyverse) 
library(magick)
library(cowplot)

statuses <- search_tweets('2019sacnas AND thinkbigdiversity OR TTLFilms', n=2000) 

statusesslim <- statuses %>% 
  filter(is_retweet == "FALSE") %>%  
  select(screen_name,  retweet_count, favorite_count, text) %>%  
  arrange(desc(favorite_count))     
head(statusesslim)

nrow(statusesslim)

statusesslim %>% 
  select(retweet_count,favorite_count) %>% 
  colSums() 
  
original <- statusesslim %>%
    group_by(screen_name) %>%
    summarize(n_tweets = n(),
              n_fav = sum(favorite_count),
              n_rt =  sum(retweet_count),
              mean_fav = round(mean(favorite_count), digits = 1),
              mean_rt = round(mean(retweet_count), digits = 1)) %>%
    filter(n_tweets >=  3) %>%
    arrange(desc(n_fav))
head(original)

mytheme <- function(){
  theme_minimal(base_size = 8) + 
  theme(panel.grid = element_blank())
}

img1 <- image_read("http://www.gradpost.ucsb.edu/images/default-source/default-album/sacnas.jpg?sfvrsn=1")
img2 <- image_read("https://pbs.twimg.com/media/EHbxW7vU0AAnWhZ?format=jpg&name=small")

img <- image_read("https://pbs.twimg.com/media/EIPMxsyWwBMdsHQ?format=jpg&name=4096x4096")
rast <- grid::rasterGrob(img, interpolate = T)

tweetsovertime <- ts_plot(statuses, "8 hour") +
  #theme(mytheme) +
  ggplot2::labs(y = "Number of tweets and retweets per 8 hours", 
                x = "Data collected from Twitter's REST API via rtweet",
                title = "Total Twitter statuses with #2019SACNAS and #ThinkBigDiversity") +
  theme_minimal(base_size = 8) + 
  theme(panel.grid = element_blank()) 

ggdraw(tweetsovertime) +
   draw_image(img1, scale = 0.3, x = -0.25, y = 0.25) +
   draw_image(img2, scale = 0.25, x = 0.3, y = 0.25) 

original %>% top_n(15, n_tweets) %>%
  ggplot() + 
  geom_bar(aes(x = reorder(screen_name, n_tweets), y = n_tweets), 
           stat = "identity", fill = "#505050") +
  geom_text(aes(label = n_tweets, y = n_tweets, x = screen_name),  
            hjust=1, size = 2, color = "#E1E9E8") +
  labs(x = NULL, y = "Tweets per user", 
       title = "Who tweeted #2019SACNAS and #ThinkBigDiversity the most?",
       caption = "Photo credit: @alexcr_1")  +
  coord_flip() +
  mytheme() + 
  annotation_custom(rast, ymin = 32.5, ymax = 80, xmin = -7)

a <- original %>% top_n(15, n_fav) %>%
  ggplot() +
  geom_bar(aes(x = reorder(screen_name, n_fav), y = n_fav),
           stat = "identity", fill = "#002855") +
  geom_text(aes(label = n_fav, y = n_fav, x = screen_name),  
            hjust=1, size = 2.5, color = "white") +
  labs(x = NULL, y =  "Total favorites", title = "Who recieved the most favorites?") +
  coord_flip() + 
  mytheme()
b <- original %>% top_n(15, n_rt) %>%
  ggplot() +
  geom_bar(aes(x = reorder(screen_name, n_rt), y = n_rt),
            stat = "identity", fill = "#DAAA00") +
  geom_text(aes(label = n_rt, y = n_rt, x = screen_name),  
            hjust=1, size = 2.5, color = "black") +
  labs(x = NULL, y  = "Total retweets", title = "The most retweets?") +
  coord_flip() + 
  mytheme()
plot_grid(a,b)

c <- original %>% top_n(15, mean_fav) %>%
  ggplot() +
  geom_bar(aes(x = reorder(screen_name, mean_fav), y = mean_fav),
           stat = "identity", fill = "#002855") +
  geom_text(aes(label = mean_fav, y = mean_fav, x = screen_name),  
            hjust=1, size = 2.5, color = "white") +
  labs(x = NULL, y = "Average favorites per tweet",
       subtitle = "Who averages the most favorites per tweet?")  +
  coord_flip() +  
  mytheme()
d <- original %>% top_n(15, mean_rt) %>%
  ggplot() +
  geom_bar(aes(x = reorder(screen_name, mean_rt), y = mean_rt),
           stat = "identity", fill = "#DAAA00") +
  geom_text(aes(label = mean_rt, y = mean_rt, x = screen_name),  
            hjust=1, size = 2.5, color = "black") +
  labs(x = NULL, y = "Average retweets per tweet", subtitle = "The most retweets per tweet?") +
  coord_flip() + 
  mytheme()
plot_grid(c,d)

foksatUCDavis <- c("alexcr_1", "Renetta_Tull", "RogersLabUCD",
                   "AnaMolinaGil3","MarkALopezPhD", "BeccaCalisi",
                   "LajoyceMboning",  "Elva_Diaz11", "ctitusbrown",
                   "UCDavisSACNAS", "raynamharris", "TTLFilms", "vdiazochoa" ,
                   "BowyerJacques", "MCalderonDeLaBS", "sociovirology",
                   "jennguerra5", "LynneArcangel", "MidoriHr", "IzaiahOrnelas",
                   "Graham_Coop", "UCDavisBiotech", "UCDavisGlobal", "UCDavisCOE",
                   "UCDavisGrad", "UCDMicrobiome", "vs_farrar", "yggdrasil13751",
                   "MedinaYarazeth", "phylogenomics", "ucdavisbiology")

retweets_total <- statuses %>% 
  filter(is_retweet == "TRUE" | is_quote == "TRUE") %>%
  select(screen_name, retweet_screen_name, retweet_count, text)   
retweets_nonucd <- statuses %>% 
  filter(is_retweet == "TRUE" | is_quote == "TRUE") %>%  
  filter(!screen_name %in% foksatUCDavis) %>%  
  select(screen_name, retweet_screen_name, retweet_count, text) 
nrow(retweets_nonucd) / nrow(retweets_total) * 100


e <- retweets_nonucd %>% 
    group_by(screen_name) %>%
    summarize(n_rt = n()) %>%
    arrange(desc(n_rt))  %>%
  head(10) %>%
  ggplot() +
  geom_bar(aes(x = reorder(screen_name, n_rt), y = n_rt),
            stat = "identity", fill = "#00acee") +
  geom_text(aes(label = n_rt, y = n_rt, x = screen_name),  
            hjust=1, size = 2.5, color = "white") +
  labs(x = "Non-UC Davis Retweeters", y  = "No. retweets", 
       title = "Which non-UC Davis tweeters amplified our message?") +
  coord_flip() + 
  mytheme()
f <- retweets_nonucd %>% 
  group_by(retweet_count,retweet_screen_name, text) %>%
  summarize(n_rt = n()) %>%
  arrange(desc(n_rt))  %>%
  mutate(who.what = paste(retweet_screen_name, text, sep = " - "),
         textslim = substr(who.what, start=1, stop=80),
         percent = round((n_rt / retweet_count *100),2)) %>%
  head(10) %>%
  ggplot() +
  geom_bar(aes(x = reorder(textslim, percent), y = percent),
            stat = "identity", fill = "#00acee") +
    geom_text(aes(label = percent, y = percent, x = textslim),  
            hjust=1, size = 2.5, color = "white") +
  labs(x = "Original Tweeter and text", y  = "% retweets by non-UC Davis Tweeters", 
       title = "Which tweets were shared most broadly?") +
  coord_flip() + 
  mytheme()
plot_grid(e,f, nrow = 2)