Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Who are my Audiences? Evolution of Target Audiences in Microblogs

1,613 views

Published on

User behavior in online social media is not static, it evolves
through the years. In Twitter, we have witnessed a maturation of its platform and its users due to endogenous and exogenous reasons. While the research using Twitter data has expanded rapidly, little work has studied the change/evolution in the Twitter ecosystem itself. In this pa-
per, we use a taxonomy of the types of tweets posted by around 4M users during 10 weeks in 2011 and 2013. We classify users according to their tweeting behavior, and nd 5 clusters for which we can associate a different dominant tweeting type. Furthermore, we observe the evolution of
users across groups between 2011 and 2013 and interesting insights such as the decrease in conversations and increase in URLs sharing. Our findings suggest that mature users evolve to adopt Twitter as a news media rather than a social network.

Published in: Social Media
  • Be the first to comment

Who are my Audiences? Evolution of Target Audiences in Microblogs

  1. 1. Who are my Audiences? Evolution of Target Audiences in Microblogs Ruth García-Gavilanes Andreas Kaltembrunner1, Diego Sáez-Trumper2 Ricardo Baeza-Yates2, Pablo Aragón1, David Laniado1 Universitat Pompeu Fabra, Barcelona Media1, Yahoo Labs2 @ruthygarcia SocInfo 2014
  2. 2. Imagined Audiences @ruthygarcia SocInfo 2014
  3. 3. Imagined Audiences @ruthygarcia SocInfo 2014
  4. 4. Contributions • Propose a taxonomy to categorize tweets by language independent features • Characterize users by the way they tweet • Analyze the evolution of imagined audiences over time @ruthygarcia SocInfo 2014
  5. 5. 8M Users who tweeted at least once during 10 weeks in 2011 (April-July) DATA @ruthygarcia SocInfo 2014
  6. 6. 8M 4.3M 2011 Users who tweeted at least once during 10 weeks (April-July) 2011 and 2013 DATA @ruthygarcia SocInfo 2014
  7. 7. Active in 2011 2013 2011 2013 Users 1,315,313 1,125,968 English Tweets 406,719,999 256,330,241 8M 4.3M 1.3M 770K 1.1M 2011 2013 2011 2013 Min 1 and max 22 tweet per working day. DATA Inactive 1 tweet per day Hyperactive 22 per day @ruthygarcia SocInfo 2014 530K 570K
  8. 8. Language Independent Approach Tweets Re-tweets (RT) Original tweets (OT) With links Without links With links Without links Mentions No Mentions Mentions No Mentions @ruthygarcia SocInfo 2014
  9. 9. Language Independent Approach No With links Without links Mentions Tweets @ruthygarcia SocInfo 2014 With links Original tweets (OT) Without links Mentions Re-tweets (RT) No Mentions Mentions % % % % % % 2011 % % % % % % 2013
  10. 10. Clusters 100% 75% 50% 25% 0% 100% 75% 50% 25% 0% 100% 75% 50% 25% 0% 100% 75% 50% 25% 0% 100% 75% 50% 25% 0% Type of tweet OT with links and mentions OT with links and no mentions OT with mentions and no links OT without links or mentions RT with links RT without links Endogenous Conversationalists Generalists Echoers Link Feeders @ruthygarcia SocInfo 2014
  11. 11. Endogenous • Social network: self-contained original posts (no mentions or links) and conversations 100% 75% 50% 25% 0% Endogenous Type of tweet OT with links and mentions OT with links and no mentions OT with mentions and no links OT without links or mentions RT with links RT without links @ruthygarcia SocInfo 2014
  12. 12. Conversationalists • Social network: emphasis on interacting with other users more than sharing self-contained ideas. Type of tweet OT with links and mentions OT with links and no mentions OT with mentions and no links OT without links or mentions RT with links RT without links Conversationalists 100% 75% 50% 25% 0% @ruthygarcia SocInfo 2014
  13. 13. Generalists • Without a distinctive type: links slightly higher. Type of tweet OT with links and mentions OT with links and no mentions OT with mentions and no links OT without links or mentions RT with links RT without links 100% 75% 50% 25% 0% Generalists @ruthygarcia SocInfo 2014
  14. 14. Echoers • News media: Forwarding other people's self-contained tweets. Type of tweet OT with links and mentions OT with links and no mentions OT with mentions and no links OT without links or mentions RT with links RT without links Echoers 100% 75% 50% 25% 0% @ruthygarcia SocInfo 2014
  15. 15. Link Feeders • Mostly tweet messages containing external links and no mentions. Type of tweet OT with links and mentions OT with links and no mentions OT with mentions and no links OT without links or mentions RT with links RT without links Link Feeders 100% 75% 50% 25% 0% @ruthygarcia SocInfo 2014
  16. 16. Sankey diagram: Users 2011 vs 2013 @ruthygarcia SocInfo 2014
  17. 17. How about users who became inactive or hyperactive? 2013 2011 Endog. Conver. Generalists Echoers L. Feeders Inactive Hyper/Bots Endog. 22.38% 5.89% 5.96% 3.56% 3.02% 58.33% 0.86% Conver. 11.33% 20.79% 7.26% 3.54% 2.41% 53.80% 0.87% Generalists 2.67% 3.88% 21.78% 2.17% 7.02% 62.07% 0.41% Echoers 9.93% 3.72% 8.31% 9.93% 3.65% 63.62% 0.84% L. Feeders 3.38% 1.47% 11.11% 1.25% 22.59% 59.45% 0.75% Inactive 6.64% 3.30% 3.12% 2.38% 2.31% 82.00% 0.26% Hyper./Bots 28.13% 17.42% 8.15% 6.48% 4.91% 26.71% 8.19% @ruthygarcia SocInfo 2014
  18. 18. CONCLUSIONS • Tendency to become “inactive” or same type of behavior over years with exception of echoers. • Decrease of conversationalists over time. • More Link Feeders and Generalists. • What is Twitter, a Social Network or a News Media? [Kwak et al., 2010] : • Mature users tend to use Twitter more as news media. @ruthygarcia SocInfo 2014
  19. 19. QUESTIONS? @ruthygarcia @ruthygarcia SocInfo 2014
  20. 20. Users per Cluster @ruthygarcia SocInfo 2014
  21. 21. Future Work • Look closely at the behavior of the inactive and hyperactive users • Lexical variation in conversations, tweets • Language vs. behavior • Change in behavior vs. change in popularity @ruthygarcia SocInfo 2014
  22. 22. K-Means k = 5 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 2 4 6 8 10 12 14 1e+05 2e+05 3e+05 4e+05 Number of Clusters Within groups sum of squares @ruthygarcia SocInfo 2014

×