doonish

 doonish

proyecto de pau gay

823790c5de8cc5bf68515352f882f1a3?s=128

Betabeers

May 13, 2012
Tweet

Transcript

  1. Online  trivia  game  to  play  with  your  friends   Pau

     Gay   @paugay   doonish   Betabeers  London  –  10/05/2012  
  2. •  Gain  points  when  asking  the  ques;ons  right  and  

    compete  to  be  the  first  into  the  ranking   •  People  can  create  ques;ons:  Community  generated   base  of  ques4ons   •  A  bit  of  story   –  2007:  First  version   –  2010:  Complete  redesign   The  Idea   Online  trivial  game  inspired  in  the     well  known  Trivial  Pursuit  
  3. Ques;on  …   Which  element  follows  the  series:   1,

     1,  2,  3,  5,  8,  13,  …  
  4. We  are  not  a  medicine,  we  are  a  vitamin  

    The  Problem   There  is  no  problem  that  we  are  solving  
  5. We  are  not  a  medicine,  we  are  a  vitamin  

    The  Problem   There  is  no  problem  that  we  are  solving   “Buen  día,  una  felicitación  para  ud  por  este  proyecto  tan  bonito  que  le  da   dignidad  al  Internet,  es  para  mi  un  gusto  estar  acá  y  que  la  casualidad  me  halla   traído.  […]”   But  some;mes  we  get  feedback  like  that:  
  6. Miguel  Florido  /  Backend  &  strategy   @miguelflorido   Crea;ve

     and  the  ideator.  Likes  to  learn  about  human  brain  and  outer  space.   SoUware  Developer  at  SoUonic  in  BCN.   Toni  Lopez  /  Backend   @picahielos   Really  hard  worker.  Likes  sports  and  the  healthy  life  and  he  is  the  king  of  the  salads.   Works  for  VisualDNA   Pau  Gay  /  Frontend  &  management   @paugay   Executor.  Likes  to  play  and  watch  football  and  strategy  games.  Working  for   VisualDNA  as  SoUware  Engineer.   The  Team   Manel  Perez  /  SEO   @do-­‐not-­‐have-­‐twi]er-­‐lol   Perfe;onist,  has  op;mized  the  applica;on.  Likes  traveling  and  the  under  the  sea  (he   met  once  The  Li]le  Mermaid).  Working  for  SoUonic.  
  7. Ques;on  …   We  have  100  birds  sibng  in  a

     tree     and  we  shoot  with  a  shotgun     and  kill  two  of  them.   How  many  birds  last  on  the  tree?  
  8.    Ini4al  approach   “Choose  the  technology  that  will  allow

     us  to  do  more  in  less  Lme”      For  that  reason:   •   Use  known  technology   •   Avoid  the  “learning”  part  at  this  stage   The  Technology  
  9. The  interes;ng  part  

  10. The  interes;ng  part   Naïve  Bayes  text  classifier   Help

     people  classify  the  ques;ons  in  a   category  depending  on  the  user  words.   See  code:   h]ps://github.com/paugay/doonish-­‐nb-­‐ classifier  
  11. The  interes;ng  part   Naïve  Bayes  text  classifier   Help

     people  classify  the  ques;ons  in  a   category  depending  on  the  user  words.   See  code:   h]ps://github.com/paugay/doonish-­‐nb-­‐ classifier   Tagging  Engine   Help  people  classify  the  ques;ons  in  a  tag   depending  on  the  using  keywords.   Controller  environment  don’t  want  to  go   crazy.  Using  included  and  excluded  keywords  
  12. The  interes;ng  part   Naïve  Bayes  text  classifier   Help

     people  classify  the  ques;ons  in  a   category  depending  on  the  user  words.   See  code:   h]ps://github.com/paugay/doonish-­‐nb-­‐ classifier   Tagging  Engine   Help  people  classify  the  ques;ons  in  a  tag   depending  on  the  using  keywords.   Controller  environment  don’t  want  to  go   crazy.  Using  included  and  excluded  keywords   Donnish  Scrapper   Robot  that  scrap  ques;ons  from  the   compe;tors:   •     TriviNet   •     Listometro   Cookie  management  to  “try  many  Lmes  unLl   get  the  right  answer”.    
  13. The  interes;ng  part   Naïve  Bayes  text  classifier   Help

     people  classify  the  ques;ons  in  a   category  depending  on  the  user  words.   See  code:   h]ps://github.com/paugay/doonish-­‐nb-­‐ classifier   Tagging  Engine   Help  people  classify  the  ques;ons  in  a  tag   depending  on  the  using  keywords.   Controller  environment  don’t  want  to  go   crazy.  Using  included  and  excluded  keywords   Wikipedia  Text   Give  some  extra  content  to  each  ques;on:   1.  Search  ques;on  on  Google   2.  Get  first  Wikipedia  entry   3.  Go  to  English  entry   4.  Get  the  first  text  snippet   5.  Translate  to  Spanish  using  Bing   Donnish  Scrapper   Robot  that  scrap  ques;ons  from  the   compe;tors:   •     TriviNet   •     Listometro   Cookie  management  to  “try  many  Lmes  unLl   get  the  right  answer”.    
  14. Ques;on  …   It  is  3.15pm  and  you  are  in

     front  of  a  old  clock.   What  is  the  angle  between  the  two     hands  of  the  clock?    
  15. Ques;on  …   It  is  3.15pm  and  you  are  in

     front  of  a  old  clock.   What  is  the  angle  between  the  two     hands  of  the  clock?    
  16. What  we  are  doing  right   •  SEO,  op;mized  landing

     page  per  ques;on     •  Layer  of  administra;on  to  approve  content   •  Keep  really  low  the  cost  of  the  project  
  17. •  Homogeny  team   •  Not  engaging  with  the  users

      •  Don’t  really  know  how  to  …   – Mone;ze   – Do  Social  Media  properly   – How  to  market  the  product   What  we  are  doing  wrong  
  18. Time  for  quesLons,  comments,  insults   The  end   Thanks

     a  lot  for  you  a]en;on   Twi]er   Facebook   twiNer.com/doonisher   facebook.com/doonisher  
  19. Interview  @  Todostartups.com   h]p://bit.ly/wXrbzM   Interview  @  betabeers.com  

    h]p://bit.ly/JzNjSA   Blog  posts  @  my  personal  blog   h]p://paugay.com/blog/category/doonish   Other  men;ons  from  the  Spanish  blogger  scene   h]p://doonish.es/apariciones   Unfortunately  content  is  in  Spanish  at  the  moment   More  about  doonish  …   Here  there  are  some  links  where  you  can  learn  more  about  doonish  and   the  project.