Twitter bio search

README.md

A friend is looking for a new tech job and asked if I knew anyone at a handful of companies. I'm (meaningfully) connected to more people who are relevant to this on Twitter than LinkedIn, so figured it would be easiest to scour people's bios. Naturally, I did this in the nerdiest way possible and automated it.

<3 @aidanfeldman

P.S. See also:

• my list of recommended job boards
• SocialRank (structured search of followers)

twitter_bio_search.rb

# usage:
# 
#   $ gem install t twitter
#   $ t authorize
#   $ ruby twitter_bio_search.rb
#   from XXX peeps:
#   
#   LEVO
#   name - @screen_name - bio
#   ...

require 'yaml'
require 'twitter'

# created via 't' gem CLI
# just grab first authentication
AUTH_INFO = YAML.load_file('/Users/afeld/.trc')['profiles'].values[0].values[0]

client = Twitter::REST::Client.new do |config|
  config.consumer_key = AUTH_INFO['consumer_key']
  config.consumer_secret = AUTH_INFO['consumer_secret']
  config.access_token = AUTH_INFO['token']
  config.access_token_secret = AUTH_INFO['secret']
end


# gather followers + following
all_ids = Set.new

client.follower_ids.each do |follower_id|
  all_ids << follower_id
end

client.friend_ids.each do |following_id|
  all_ids << following_id
end

puts "from #{all_ids.size} peeps:\n\n"


TERM_REGEX = /birchbox|bonobos|codecademy|etsy|foursquare|gilt|google|levo|okcupid|seamless|spotify|squarespace|streeteasy|tumblr|vimeo|warby/i

users_by_terms = {}

# can only retrieve 100 at a time
all_ids.each_slice(100) do |some_ids|
  client.users(*some_ids, include_entities: false).each do |user|
    desc = user.description
    if desc
      desc.scan(TERM_REGEX) do |term|
        term = term.downcase
        users_by_terms[term] ||= Set.new
        users_by_terms[term] << user
      end
    end
  end
end


users_by_terms.sort.each do |term, users|
  puts term.upcase

  users.each do |user|
    desc = user.description

    if desc
      # compress whitespace
      desc = desc.gsub(/\s+/, ' ')
    end

    puts "https://twitter.com/#{user.screen_name} – #{user.name} – #{desc}"
  end

  puts "\n"
end

I have created a fork at https://gist.github.com/almereyda/e762a3d50c6818090c46ecad594dcadc which also searches the location field.