First letter of the quote tweet author last name who replied to https://twitter.com/rafaelotinoco/status/1329622292507267073 – https://twitter.com/ikashnitsky/status/1330233541674885123

twitter-b-census.R

#===============================================================================
# 2020-11-21 -- twitter
# B census in replies to the joke tweet
# https://twitter.com/rafaelotinoco/status/1329622292507267073
# Ilya Kashnitsky, ilya.kashnitsky@gmail.com
#=============================================================================== 

# analyse the first surname letter of those replied to the meme
# the prevalence of simila jokes from B people was so obvious

library(tidyverse)
library(magrittr)
library(stringi)
library(rtweet)
library(ggdark)
library(hrbrthemes)

# for replicability the export is done at this time
##------ Sat Nov 21 19:10:05 2020 ------##
df <- search_tweets("url:1329622292507267073", n = 2000)

save(df, file = "tweets.rda")

# filter quoting tweets
df_quo <- df %>% filter(is_quote == T)

# get the info in the quoting authors names
quo_auth <- df_quo %>% users_data()

# tally first letter of the surnames
df_name <- quo_auth %>% 
    select(name) %>% 
    mutate(
        name = name %>% str_to_lower() %>% 
            stri_enc_toascii() %>% 
            # remove text in parentheses
            str_remove("\\s*\\([^\\)]+\\)") %>% 
            # remove hasgtags
            str_remove("(?<=^|\\s)#[^\\s]+") %>% 
            #remove "jr"
            str_remove(" jr")
    ) %>% 
    # remove the common ", phd"
    separate(name, sep = ",", into = c("one", "two")) %>% 
    select("one") %>% 
    # get the last word in the name string ~surname
    mutate(name = one %>% str_extract("[^ ]+$")) %>% 
    # if NA coppy from name
    transmute(surname = case_when(is.na(name)~one, TRUE~name)) %>% 
    # GET THE FIRST LETTER OF THE SURNAME
    mutate(first = surname %>% str_sub(1, 1)) %>% 
    # calculate letters freq
    group_by(first) %>% 
    summarise(n = n()) %>% 
    ungroup()

# get the dataset for plotting
df_plot <- tibble(LETTERS) %>% 
    mutate(first = LETTERS %>% str_to_lower) %>% 
    left_join(df_name) %>% 
    mutate(prop = n %>% prop.table())

# visualize
df_plot %>% 
    ggplot(aes(LETTERS, prop))+
    geom_col(color = NA, fill= "orange", width = .75)+
    scale_y_percent()+
    labs(
        title = "% of quote tweets",
        x = "First letter of the quote tweet author last name",
        y = NULL
    )+
    dark_theme_minimal(base_family = "Roboto Slab")+
    theme(
        axis.text.x = element_text(
            size = 14, face = 2, colour = c("white", "orange", rep("white", 24))
        )
    )+
    annotate(
        "text", x = 4, y = .3, hjust = 0, vjust = 1, 
        color = "orange", size = 7, family = "Roboto Slab",
        label = 'The Twitter census of (B)est jokes'
    )+
    annotate(
        "text", x = 4, y = .23, hjust = 0, vjust = 1, 
        color = "white", size = 5, family = "Roboto Slab",
        label = "A good joke is always unexpected, isn't it?"
    )

ggsave(filename = "out.png", width = 6, height = 3.375)