For Dr. Emma Hodcroft

get_twitter_handles_from_google.R

library(rvest)
library(rio)
library(dplyr)
library(xml2)

# If your spreadsheet is named "data.xlsx" and the column with submitter names is named "submitting_lab"
data <- rio::import("data.xlsx") %>%
  dplyr::pull(submitting_lab) %>%
  unique()


# This function gets the 3 top possible Twitter handles from a Google search. It sleeps for 3 seconds before returning results not to pound Google if running on multiple items
get_google_top_links <- function(submitter) {
  submitter_encoded <- gsub(" ", "+", submitter, fixed = TRUE)
  url <- URLencode(paste0("https://www.google.com/search?q=site%3Atwitter.com+", submitter_encoded))
  my_html <- xml2::read_html(url)
  titles <- rvest::html_nodes(my_html, xpath = "//div/div/div/a/div[not(div)]") %>% 
    rvest::html_text() 
  titles <- titles[1:3]
  titles <- gsub("twitter.com › ", "", titles, fixed = TRUE)
  titles <- gsub(" › status.*?$", "", titles)
  mydf <- data.frame(Submitter = submitter, Google1 = paste0("@", titles[1]), Google2 = paste0("@", titles[2]), Google3 = paste0("@", titles[3]))
  Sys.sleep(3)
  return(mydf)
}


# test for 10 items
my_results <- purrr::map_df(data[1:10], get_google_top_links)

# OPTIONAL: Get HTML table of results so you can click to see whether you want those Twitter handles


library(DT)
library(htmltools)
get_clickable_url <- function(handle) {
  handle1 <- gsub("@", "", handle, fixed = TRUE)
  url <- paste0("<a href='https://twitter.com/", handle1, "'>", handle, "</a>")
  return(url)
}

Google1links <- purrr::map_chr(my_results$Google1, get_clickable_url)
Google2links <- purrr::map_chr(my_results$Google2, get_clickable_url)
Google3links <- purrr::map_chr(my_results$Google3, get_clickable_url)

my_results_table_data <- my_results %>%
  mutate(
     Google1 = Google1links,
     Google2 = Google2links,
     Google3 = Google2links
  )
  
  
DT::datatable(my_results_table_data, escape = FALSE)