Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Social Media Mining

Social Media Mining

Vinti Maheshwari

November 21, 2014
Tweet

More Decks by Vinti Maheshwari

Other Decks in Technology

Transcript

  1. CS550-­‐  Project1   Data  Mining  On  Social  Media  Advent,  The

     Pros  and  Cons     Fall-­‐2014     Prof  :  Alex  Wu    
  2. Social  Media   Social  media  is  digital  content  and  interac8on

     that  is  created  by  and  between  people.  
  3. What  is  Data  Mining   Generally,  data  mining   (some8mes

     called  data  or   knowledge  discovery)  is  the   process  of  analyzing  data   from  different  perspec8ves   and  summarizing  it  into   useful  informa8on.     Informa8on  that  can  be  used   to  increase  revenue,  cuts   costs.  Or  both.    
  4. Social  Media  Mining   •  What  is  social  media?  

      •  Why  we  need  it?   •  How  we  perform  social  media  analy8cs?    
  5. What  is  Social  Media  Mining?   •  Social  media  analy8cs

     is  process  of  collec8on  data   from  the  social  sites  like  Facebook,  linked  in,  twiQer  ,   Google+  and  so  on…  Most  of  the  professionals  are  on   linked  in.   •  People  are  pos8ng  informa8on  con8nuously.  Blogs  are   also  big  source  of  data,  many  of  you  are  on  bloggers   also.   •  Collec8ng  all  the  informa8on  available  on  social  media.   What  informa8on  about  par8cular  product  or  brand.   •  Fetching    +  analysis  (  try  to  idenGfy  insights)  -­‐-­‐-­‐>   taking  business  decision   •   whole  process  is  social  media  analy8cs.  
  6. Social  Media  AnalyGcs  -­‐  Why?     1.  MarkeGng:  target

      2.  Customer  Service:  sen8ment  part  (to  find  out   are  customer  unhappy,  complaining  about   product,  way  to  improve)   3.  Brand  ReputaGon:  its  very  difficult  to  build  a   brand  and  very  easy  to  fall  down.  People  are   not  damaging.  So  they  take  correc8ve  ac8on   at  8me.       Social  Media  Monitoring   Monitoring  is  the  huge  part  of  social  media  analy8cs.  
  7. Measures   •  QuanGtaGve  Measures    Is  there  any  buzz

     around  the  topic,  it  may  be    posi8ve  or  nega8ve   •  QualitaGve  Measures   – We  go  deeper  under  QuanGtaGve  measure   analysis  to  find  out  how  many  8mes  topic   men8oned  posi8ves  and  how  many  nega8ve.  
  8. Methodology   •  Set  Business  Goal   •  Set  KPI(s)

     (key   performance   Indicator)   •  Gather  Social  Media   Data   •  Perform  Analysis   •  Measure  &   Feedback  
  9. Social  Media  Mining-­‐  How?     1.  Readily  available  tools:

     free,  paid     2.  Customer  Build  -­‐  Using  R,  Text  Mining,  Web   UI  
  10. Tools   •  33Across  hTp://www.33across.com/   —  it  helps  brands

     get  a  more  complete  picture   of  their  online  audience  by  analyzing  how  they   interact  with  the  brand’s  social  media   interac8ons.  That’s  good  for  both  companies   and  their  customers:  33Across  helps  companies   figure  out  the  best  places  to  focus  their   markeGng  efforts,  which  leads  to  more  relevant   ads  for  consumers.  
  11. Tools   Brandwatch  hTp://www.brandwatch.com/       •  dog  listening

     For  interna8onal  corpora8ons  with   interna8onal  reach,  it  supports  25  languages.   •  Like  most  other  analy8cs  suites,  Brandwatch   monitors  all  of  the  conversa8ons  across  various   social  networks.     •  The  company  claims  that  its  algorithm  cuts   through  social  spam  to  help  find  the  conversa8ons   that  really  maQer.  It’s  known  for  having  some  of   the  best  data  gathering  and  parsing  tools     •  social  media  management  company  Hootsuite   partners  with  Brandwatch  to  enable  its  customer   to  benefit  from  Brandwatch’s  70-­‐million-­‐plus  data   data  sources.  
  12. Tools   Google  AnalyGcs    hTp://www.google.com/analyGcs/     •  Great

     resource  for  social  media  analy8cs  on  your   website.     •  Last  year  Google  added  social  reports  to  analy8cs,   which  you  can  use  to  determine  the  conversion  value   of  visitors  from  social  sites  as  well  as  see  how  visitors   from  different  social  sites  behave  on  your  site.     •  It  also  has  an  “Ac8vity  Stream”  that  shows  in  real  8me   how  people  are  talking  about  your  site  on  social   networks.  
  13. Case  Study:  TwiQer   Steps  Involved   •  Crea8ng  a

     TwiQer  API  Connec8on   •  Searching  for  Tweets   •  Analyzing  the  140  characters.   – Extrac8ng  tweet  en88es   – Analyzing  tweets  en8ty  with  frequency   – Visualizing  frequency  data  in  re  tweets    
  14. Exploring Twitter website: acquiring and analyzing Twitter data. Tweet:  @ptwobrussell

     is   wri8ng  @SocialWebMining,   2nd  Ed.  from  his  home   office  in  Franklin,  TN.  Be   #social:  hQp://on.p.me/ 16WJAf9     EnGGes:  ptwobrussell,   SocialWebMining  ,  Franklin,   TN,  hQp://on.p.me/ 16WJAf9        
  15. Adver8sements  on  Facebook   Facebook  know   our  age  and

      gender  and  on  the   basis  of  this  data,   they  gives  add  on   sidebar   Facebook  adds  for  woman   Facebook  adds  for  man   Facebook  is  blessed   enough  to  have   user’s  personal   data  as  we   ourselves  provide  it   willingly.    
  16. Conclusion   •  Business  can  analyze  their  customer   behavior

     and  trends.   •  Customer  can  get  help  in  decision   making  what  to  buy.   •  Helps  in  adver8sing   ü  Security  /  privacy  issues   ü  Many  companies  watching  you   what  you  are  browsing  ,  it’s  not   safe  always.      
  17. References   •  hQp://shop.oreilly.com/product/ 0636920030195.do   •  hQps://chrome.google.com/webstore/detail/ ghostery/mlomiejdvolichcflejclcbmpeaniij?hl=en  

    •  hQps://www.youtube.com/watch?v=-­‐ ro9328gS_U     •  hQps://www.youtube.com/watch? v=g7Xbbx67ObY&list=PL0174E49C0E0DD5C8&ind ex=12