Function to automate generation of RDS, simple CSV and plot using rtweet

extract_tweets_function.R

# Function to automate generation of RDS, simple CSV and plot using rtweet
# Matt Dray
# March 2018

# Purpose: create an RDS, simplified CSV and plot of tweets containing search term
# from rtweet::get_tweets function and save to to folder with unique descriptive
# name related to search term. Assumes you have an 'output' folder in your home
# directory to store these files. Assumes you've already sorted out a twitter
# token as per http://rtweet.info/articles/auth.html

extract_tweets <- function(search_term = NULL, ...){
  
  # packages
  
  library(rtweet)  # for getting tweets
  library(ggplot2)  # for plotting
  
  # create folder to hold outputs
  
  main_dir <- getwd()  # home directory hopefully
  
  file_name <- paste0(  # create format for file name
    format(Sys.time(), "%y%m%d"),  # today's date in form 180314
    "_",
    gsub("[[:punct:][:blank:]]+", "", search_term)  # search term no punctuation
  )
  
  dir.create(file.path(main_dir, "outputs", file_name))  # create folder
  
  # perform search
  
  temp_search <- rtweet::search_tweets(
    q = search_term,
    n = ...,
    include_rts = FALSE
  )
  
  # save full dataset as RDS
  
  saveRDS(
    object = temp_search,
    file = paste0(file.path(main_dir, "outputs", file_name, file_name), ".RDS")
  )
  
  # simplify object
  
  temp_simple <- temp_search[, c("screen_name", "created_at", "text")]
  
  # write simplified object to csv
  
  write.csv(
    x = temp_simple,
    file = paste0(file.path(main_dir, "outputs", file_name, file_name), ".csv")
  )
  
  # create simple plot
  
  temp_plot <- rtweet::ts_plot(data = temp_search, by = "3 hours") +
    ggplot2::theme_minimal() +
    ggplot2::theme(plot.title = ggplot2::element_text(face = "bold")) +
    ggplot2::labs(
      x = NULL, y = NULL,
      title = paste0(
        "Frequency of \"",
        gsub(":[punct]:", "", search_term),
        "\" occurrences in tweets"
      ),
      subtitle = "Aggregated Twitter status (tweet) counts",
      caption = "\nSource: Data collected from Twitter's REST API via rtweet"
    )
  
  # save simple plot
  
  ggplot2::ggsave(
    file = paste0(file.path(main_dir, "outputs", file_name, file_name), ".png"),
    plot = temp_plot,
    width = 10,
    height = 7
  )
  
}