project goal.

This project is a product of a course titled Computational Methods in the Humanities, an honors course offered at the University of Pittsburgh. The course is taught by Dr. David Birnbaum. The project participants include Devon Broglie (’15), Kimaya Padgaonkar (’15), and Sung Wi (’13).

Our goal was to identify common themes among parody accounts in the attempt to determine the draw of following a parody account. We were analyzing the content of tweets by the topic they relate to (themes) and the twitter elements they contain (hashtags, mentions, media, etc.). We were also collecting daily statistics for each accounts’ tweets, following, and followers to see how these change over time.

twitter lingo.

Twitter is an online social networking tool in which users post 140 character updates of what is going on in their lives along with links to things they think are interesting, funny, or useful to their followers (“following” being essentially what “friending” is on other sites). People use twitter in many ways, some as a newsfeed by following prominent people or networks, some as a pseudo-chatroom by limiting their followers and whom they follow to close friends and family, and some as a microblog for updating people about the work they are doing and their personal lives.

general definitions.

  • A tweet is a 140 character or less post.
  • A follower, is a user who chooses to “follow” your account (your tweets will appear in their feed).
  • The accounts a user chooses to follow are grouped in the “following” category.

content definitions.

  • A mention is when a tweet includes another user’s twitter handle (username). This is designated by an ‘@’ sign preceding the username.
  • A hashtag is a trend that relates to the tweet content. This is designated by a ‘#’ preceding the trend.
  • A retweet is when one user chooses to “share” another user’s tweet. It is essentially quoting the other user.
  • Other media, such as pictures, video, web links, etc. appear as hyperlinks in the tweet.

project specific definitions.

  • Verification is a way for twitter to identify actual celebrity accounts. These are marked by a checkmark in a blue bubble next to the username.
  • A parody account is one that poses as a real celebrity account, though it is not verified and includes a disclaimer (that it is not real or it is a parody) in the account bio.

Data Collection

The data was collected over a two week period form March 12, 2013 to March 26, 2013. Statistics were collected for six parody accounts and their corresponding real celebrity accounts. The statistics included daily numbers of tweets, followers, following through this two week period. Tweets were collected from the six parody accounts everyday. These tweets were then tagged in XML in order to use them for analysis.

XML Tagging of Tweets

The tweets were each individually tagged with a theme attribute. The themes are listed below with their tag name in parentheses as well as examples or descriptions of tweets that would fit the theme.

  • Social media advertisements (media)
    • Follow back
    • Encouraging Facebook page likes
  • Third-party ads (ad)
    • Sponsor
    • Links to sites
  • Social events (social)
    • Includes partying
  • Daily life (daily)
    • “went to the gym”
  • Milestones (mile)
    • Major life events
    • Birthday, awards, shows, tours etc.
  • Relationship advice (relation)
    • Love related
    • Any relationship
  • Wisdom (wisdom)
    • Advice
    • Proverbs
  • Gossip (gossip)
    • Rumors
  • Conversations (convo)
    • With other accounts through mentions
  • Jokes/humor (joke)
    • Includes memes
  • General thoughts (general)
    • Diary like
    • Feelings not related to other themes
  • Fans (fan)
    • Retweeting fans
    • Appreciating fans

Elements within tweets were also tagged. The elements are listed below with their tag name in parentheses.

  • Picture link (pic)
  • Video link (vid)
  • Website link (link)
  • Mention (mention)
  • Hash tag (hash)
  • Emoticons (emo)

project findings.

Our data analysis gave us clear findings. Based on observation of the graphs showing trends of themes and daily statistics, we were able to determine that there is a correlation between parody account tweets and their followers. We were able to measure the success of an account based on a mathematical formula, called the leadership ratio. We defined this ratio as the change in followers over the change in following over the two week period (March 12, 2013 to March 26, 2013) that the tweets were collected. Individual analysis for each parody account is located under the celebrities tab.