Mapping the South African Twittersphere

4,942 views

Published on

This deck briefly outlines the work we did mapping the South African Twittersphere for the 2012 SAMRA conference, including some analyses we did based on the structure of the network. Specifically, we identified people with the potential for influence based on their betweeness centrality and Authority (HITS). In addition, we also used a modularity algorithm to identify 5 clearly distinct communities within the graph. The results are for interest-sake only and should be interpreted within the limitations of the data."

1 Comment
7 Likes
Statistics
Notes
  • Hi,

    Your presentation is absolutely amazing and the images for structural networks are really vivid and interesting.
    My first thought was that the twitter user name is SAP resident (I’m interested in CRM, so it was my association and I even wanted to follow it), but then I realized that it's actually South African President's twitter name. :)

    I would have definitely included you either in Top Ten Technology Authority or Top Ten proudly SA&DA Authority.

    Thanks for your creative science!
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total views
4,942
On SlideShare
0
From Embeds
0
Number of Embeds
266
Actions
Shares
0
Downloads
65
Comments
1
Likes
7
Embeds 0
No embeds

No notes for slide

Mapping the South African Twittersphere

  1. Mapping the South AfricanPolitical Twittersphere
  2. Who are the popular Who is influencing What is the nature What sub- and/or influential whom within a of their influence i.e. communities exist individuals in my specific topic of how are they within my online network? interest to me? influencing people? community? What topics are What are people saying What do these sub- Who is sharing about my brand, communities believe people talking political party, charity, and how do they relate my content about? music, etc.? to my brand/ online? organisation/etc? What are the Are they saying How far is word of How is my emerging topics mostly positive or my brand or messaging being and trends? negative things? negative publicity shared online? being spread?Relevance?
  3. The South African Political Twittersphere
  4. A paradigm shift… Participants not respondents Market (CASRO) researchers “Private gardens” vs. public spaces misunderstand social media Advocating an environment where traditional research can’t compete with research… 3rd party developersSource: Barber, Michael, 2011
  5. Actual tweetsTwo types of dataTwitter user data
  6. Tweets data summary Number of tweets collected: @helenzille = 28,500 @PresidencyZA = 3,500 @SAPresident = 2,900 Time period: Sept 2011 – March 2012
  7. Target nodes Source nodes Collected 64,357 Cleaned 58,349 Distance <= 4 37,695 2.5 million Distance <= 3 20,460 Distance <= 2 12,753 Distance <= 1 4,479 Distance = 0 3 Time period: April 2012Twitter user data summary
  8. 3.0 2.0 1.0 Ego
  9. 2.0 1.5 1.0 HelenZille 2.0 1.5 1.0 SAPresident PresidencyZA 2.0 1.5 1.0
  10. Topic modelling HashtagsTypes of analyses Structural network Betweeness centrality Authority (HITS) Community detection Influence networks Image sources: Knowledge Matters. http://www.durantlaw.info/category/miscellaneous/strategy Gutiérrez-Pérez, JA, et al. 2011. Application of graph-spectral methods in the vulnerability assessment of water supply networks
  11. Actual tweetsTwo types of data
  12. @SAPresident + @PresidencyZA
  13. @helenzille
  14. Two types of dataTwitter user data
  15. The South African politicalTwittersphere
  16. Structural network by follower count
  17. Top 10 byfollower count
  18. Structural network bybetweeness centrality
  19. Top 10 bybetweeness centrality
  20. Structural network by Authority
  21. Top 10 byAuthority
  22. Black influentials News mediaTechnology Proudly SA & DA Sportsmen & celebrities Community detection Louvain
  23. Community detectionLouvain
  24. Top 10 sportsmen & celebritiesFollower countHelen Zille Bryan Habana John Smit Albie Morkel Mark Boucher157,642 100,141 84,529 94,593 84,564Cricket South South African Trevor Immelman Steve Hofmeyr 94.7 HighveldAfrica 61,536 Rugby 60,280 53,777 39,351 Radio 37,554
  25. Top 10 sportsmen & celebritiesBetweeness centralityRob van Vuuren Arno Carstens Cape Talk 567 Alex van Tonder Sam Wilson16,610 11,562 11,685 10,065 6,756Karen Zoid Lead SA Jeannie D Creative Cape The Foodie (David)14,410 18,786 38,858 Town 6,259 5,219
  26. Top 10 news media Follower countYaseen Theba Mail & Guardian South African BBC Africa Redi Tlhabi77,191 57,320 Presidency 53,413 49,425 45,866SA Breaking News City Press allAfrica.com Mandy Wiener Evita Bezuidenhout38,487 36,376 32,788 31,296 30,139
  27. Top 10 news media Betweeness centralityMandy Wiener Sam Mkokeli Ferial Haffajee Mail & Guardian Robyn Clark31,296 3,033 29,530 57,320 1,598Sipho Hlongwane US South African Verashni Pillay The Big Issue Stephen Grootes9,093 Embassy 17,803 8,304 4,363 19,708
  28. Top 10 technologyFollower countAki Anastasiou Jaco van Wyk Simon Dingle Dave Duarte Toby Shapshak28,063 20,087 15,386 9,563 9,073Finance24 Mxit Matthew Buckland Pieter Uys Duncan McLeod8,184 7,651 7,033 6,654 6,194
  29. Top 10 technologyBetweeness centralityMatthew Buckland Dave Duarte Aki Anastasiou Raoul de Jongh Toby Shapshak7,033 9,563 28,063 4,133 9,073Saul Kropman Mike Sharman Uno de Waal Pete Flynn Cathryn Reece2,530 4,680 2,469 1,008 2,302
  30. Top 10 black influentials Follower countBonang Matheba Julius Malema Dineo Ranaka Jacob Zuma Black Coffee231,696 (unconfirmed) 192,482 142,745 136,403 97,302David Kau Terry Pheto Pabi Moloi Claire Mawisa Kuli Roberts83,169 64,117 66,841 54,562 57,352
  31. Top 10 black influentials Betweeness centralityKhaya Dlanga Simphiwe Dana TimesLIVE Julius Malema Xolisa Dyeshana46,101 33,986 52,706 (unconfirmed) 192,482 4,747Jason Von Berg David Kau Zama Ndlovu Bonang Matheba Mvelase Peppetta4,445 83,169 6,718 231,696 2,804
  32. Top 10 proudly SA & DAFollower countLindiwe Mazibuko SA The Good News Brand South Africa SouthAfrica.info City of Cape Town30,460 12,483 8,549 8,575 7,725Ryan Coetzee Mmusi Maimane Tim Harris DA Youth Andrew Boraine3,528 4,917 2,792 3,239 2,214
  33. Top 10 proudly SA & DABetweeness centralityLindiwe Mazibuko Ryan Coetzee Climate Smart Tim Harris Mmusi Maimane30,460 3,528 Cape Town 867 2,792 4,917Solly Malatsi Gareth van Mbali Ntuli Cape Town Phumzile Van853 Onselen 2,085 1,903 Green Map 1,408 Damme 1,588
  34. Influence is contextual
  35. Actual tweetsTwo types of data
  36. #ANCYL#ANCYL #malema #ANCYLmarch | 143 tweets
  37. #POIB#POIB #POSIB #stopthesecrecybill #secrecybill#BlackTuesday | 654 tweets Interesting: @PatriciaDeLille: • High presence (mentions) • Low influence (interactions)
  38. @hopeleighm @PresidencyZA @SAPresident@helenzille and membersof the Democratic Alliance
  39. A new kind of research Complex systems Influence and dynamics Public data MR = computer science CONCLUSIONSParadigm shift Representative what!? Participants, not respondents 3rd party disruptive innovation
  40. Thank you@socialphysicist
  41. Appendices
  42. Top 10 sportsmen & celebritiesAuthorityHelen Zille Rob van Vuuren John Smit Arno Carstens Cape Talk 567Bryan Habana Alex van Tonder Sam Wilson Karen Zoid Vanessa Raphaely
  43. Top 10 news media Authority South AfricanMail & Guardian Ferial Haffajee Mandy Wiener Stephen Grootes PresidencyEvita Bezuidenhout Redi Tlhabi Zapiro Max du Preez Philip de Wet
  44. Top 10 technologyAuthorityAki Anastasiou Matthew Buckland Toby Shapshak Dave Duarte Simon DingleDuncan McLeod Memeburn Cherryflava Media Pieter Uys Mike Sharman
  45. Top 10 black influentials Authority Julius MalemaJacob Zuma TimesLIVE Khaya Dlanga Leanne Manas (unconfirmed)Riaad Moosa Zakes Mda David Kau Simphiwe Dana Pabi Moloi
  46. Top 10 proudly SA & DAAuthorityLindiwe Mazibuko Ryan Coetzee SA The Good News City of Cape Town Mmusi Maimane Gareth vanTim Harris David Maynier Andrew Boraine Gareth Morgan Onselen
  47. Topicmodelling

×