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

2012 ConvergeSE: Exploring Human Identity Through Personalization and Data Mining

2012 ConvergeSE: Exploring Human Identity Through Personalization and Data Mining

This workshop dives into the depths of how a person can be identified online, allowing for the delivery of highly personalized experiences. This will cover a few topics, including:
Building the identity foundation using authentication systems like BrowserID, PayPal Access and Facebook Connect.
Understanding how cultural identity concepts like tribalism play into how people group themselves innately online.
Building personality and interest profiles for users by tracking actions using keyword density scraping and categorization.

Jonathan LeBlanc

April 24, 2012
Tweet

More Decks by Jonathan LeBlanc

Other Decks in Technology

Transcript

  1. Through  Personaliza0on  and  Data  Mining   Exploring  Human  Iden0ty  

    Jonathan  LeBlanc   Developer  Evangelist:  X.commerce   Twi;er:  @jcleblanc   E-­‐Mail:  [email protected]   Github:  github.com/jcleblanc  
  2. h;p://www.x.com   h;p://slideshare.net/jcleblanc   The  FoundaGons  of  Human  IdenGty  

    Tribalism  and  Social  Grouping   Experimental  IdenGty  Methods   The  Big  Bag  of  Social  IdenGty  Fail   What  We’re  Going  to  Cover  
  3. h;p://www.x.com   h;p://slideshare.net/jcleblanc   The  Founda0ons  of  Human  Iden0ty  

    Tribalism  and  Social  Grouping   Experimental  IdenGty  Methods   The  Big  Bag  of  Social  IdenGty  Fail   What  We’re  Going  to  Cover  
  4. h;p://www.x.com   h;p://slideshare.net/jcleblanc   OAuth  (1.0a  +  2.0)   PayPal

     Access,  Facebook,  Twi;er     OpenID  (…and  the  upcoming  OpenID  Connect)   PayPal  Access,  Google,  Yahoo!     BrowserID   Mozilla   Human  Iden0ty:  Open  Iden0ty  Programming  
  5. h;p://www.x.com   h;p://slideshare.net/jcleblanc   There  are  a  few  common  op0ons

      Human  Iden0ty:  Tracking  Anonymous  Users   Tracking  Cookie   Local  Storage  
  6. h;p://www.x.com   h;p://slideshare.net/jcleblanc   •  On  each  page  visited,  track

     the  URL   •  HTML5  Local  Storage  as  primary  storage   •  Cookies  as  secondary  storage   Human  Iden0ty:  Tracking  Anonymous  Users   Program  Overview  
  7. Tracking  Anonymous  Users  with  Local  Storage   var  storeName  =

     "visited";   if  (typeof(localStorage)  ==  'undefined'  )  {          //Local  Storage  Not  Available   }  else  {          try  {                  var  sites  =  localStorage.getItem(storeName);                  sites  =  (sites  ===  null)  ?  window.loca0on  :  sites  +  window.loca0on;                    localStorage.setItem(storeName,  sites  +  "|");            }  catch  (e)  {                  if  (e  ==  QUOTA_EXCEEDED_ERR)  {                          //quota  exceeded                  }          }   }  
  8. Tracking  Anonymous  Users  with  Cookies   func0on  readCookie(name)  {  

           var  nameEQ  =  name  +  "=";          var  ca  =  document.cookie.split(';');          for  (var  i  =  0;  i  <  ca.length;  i++)  {                  var  c  =  ca[i];                  while  (c.charAt(0)  ==  '  '){  c  =  c.substring(1,  c.length)  };                  if  (c.indexOf(nameEQ)  ==  0){                            return  c.substring(nameEQ.length,  c.length);                    }          }          return  null;   }  
  9. var  storeName  =  "visited";   if  (typeof(localStorage)  ==  "undefined"  )

     {          var  cookieVal  =  readCookie(storeName);          var  value  =  ((cookieVal  ===  null)  ?  window.locaGon  :  cookieVal                                                  +  window.locaGon);                      var  days  =  1;          var  date  =  new  Date();          date.setTime(date.getTime()  +  (days*24*60*60*1000));          var  expires  =  ";  expires="  +  date.toGMTString();          document.cookie  =  storeName  +  "="  +  value  +  "|"                                                                            +  expires  +  ";  path=/";   }  else  {          //Use  Local  Storage   }   Tracking  Anonymous  Users  with  Cookies  
  10. h;p://www.x.com   h;p://slideshare.net/jcleblanc   •  Remove  oldest  results  when  storage

     fills   •  Build  categorizaGon  mapping  prior  to   storage  to  save  space  (more  on  this  later)   Human  Iden0ty:  Tracking  Anonymous  Users   Next  Steps  /  Improvements  
  11. h;p://www.x.com   h;p://slideshare.net/jcleblanc   Human  Iden0ty:  Real  Iden0ty  Sources  

    Social  (perceived)   Concrete  (true)       Sources  of  Real  Iden0ty  
  12. h;p://www.x.com   h;p://slideshare.net/jcleblanc   PayPal  Access:  OAuth  2  +  Commerce

        Seamless  Checkout   Prospect  Scores   RecommendaGons  
  13. PayPal  Access:  The  Common  Code     <?php   define('KEY',

     'YOUR  APPLICATION  ID');   define('SECRET',  'YOUR  APPLICATION  SECRET');     define('CALLBACK_URL',  'YOUR  CALLBACK  PATH  -­‐  TO  COMPLETE.PHP');   define('AUTH_ENDPOINT',  'h;ps://idenGty.x.com/xidenGty/resources/authorize');   define('TOKEN_ENDPOINT',  'h;ps://idenGty.x.com/xidenGty/oauthtokenservice');   define('USER_ENDPOINT',  'h;ps://idenGty.x.com/xidenGty/resources/profile/me');     func0on  run_curl($url,  $method  =  'GET',  $postvals  =  null){  ...  }   ?>  
  14. PayPal  Access:  Forwarding  for  Login   <?php   require_once  "common.php";

        $auth_url  =  sprinb(          "%s?scope=%s&response_type=code&redirect_uri=%s&client_id=%s",          AUTHORIZATION_ENDPOINT,          urlencode("h;ps://idenGty.x.com/xidenGty/resources/profile/me"),          urlencode(CALLBACK_URL),          KEY);     //forward  user  to  PayPal  auth  page   header("LocaGon:  $auth_url");   ?>    
  15. PayPal  Access:  Obtaining  the  Access  Token   <?php   require_once

     "common.php";     //capture  code  from  auth   $code  =  $_GET["code"];     //construct  POST  object  for  access  token  fetch  request   $postvals  =  sprinb("client_id=%s&client_secret= %s&grant_type=authorizaGon_code&code=%s&redirect_uri=%s",            KEY,            SECRET,          $code,          urlencode(CALLBACK_URL));     //get  JSON  access  token  object   $token  =  json_decode(run_curl(ACCESS_TOKEN_ENDPOINT,  'POST',   $postvals));    
  16. PayPal  Access:  Using  the  Access  Token   //construct  URI  to

     fetch  profile  informaGon  for  current  user   $profile_url  =  sprinb("%s?oauth_token=%s",   PROFILE_ENDPOINT,  $token-­‐>access_token);     //fetch  profile  of  current  user   $profile  =  run_curl($profile_url);     var_dump($profile);   ?>    
  17. h;p://www.x.com   h;p://slideshare.net/jcleblanc   The  FoundaGons  of  Human  IdenGty  

    Tribalism  and  Social  Grouping   Experimental  IdenGty  Methods   The  Big  Bag  of  Social  IdenGty  Fail   What  We’re  Going  to  Cover  
  18. h;p://www.x.com   h;p://slideshare.net/jcleblanc   Tribalism  started  as  a  way  to

     keep  us  safe     …it  has  lead  to  some  horrible  parts  of  history     but  it  is  also  a  foundaGon  of  many  of  our  social   relaGonships       Social  Grouping:  Founda0on  in  Tribalism  
  19. h;p://www.x.com   h;p://slideshare.net/jcleblanc   •  Use  all  URLs  from  the

     previous  program.   •  Obtain  content  category  for  page.   •  Categorize  user  interest.     Social  Grouping:  Group  Programming  Primer   Program  Overview  
  20. h;p://www.x.com   h;p://slideshare.net/jcleblanc   The  FoundaGons  of  Human  IdenGty  

    Tribalism  and  Social  Grouping   Experimental  Iden0ty  Methods   The  Big  Bag  of  Social  IdenGty  Fail   What  We’re  Going  to  Cover  
  21. h;p://www.x.com   h;p://slideshare.net/jcleblanc   Step  1:  Perform  Discovery    

    curl  h;ps://gmail.com/.well-­‐known/host-­‐meta       Experimental  Iden0ty:  WebFinger  
  22. h;p://www.x.com   h;p://slideshare.net/jcleblanc   Experimental  Iden0ty:  WebFinger   <XRD  xmlns='h;p://docs.oasis.open.org/ns/xri/xrd-­‐1.0'

                 xmlns:hm='h;p://host-­‐meta.net/xrd/1.0'>          <hm:Host  xmlns='h;p://host-­‐meta.net/xrd/1.0'>gmail.com          </hm:Host>          <Link  rel='lrdd'                    template='h;p://www.google.com/s2/webfinger/?q={uri}'>                  <Title>Resource  Descriptor</Title>          </Link>   </XRD>  
  23. h;p://www.x.com   h;p://slideshare.net/jcleblanc   Step  2:  Collect  User  Data  

      curl  h;p://www.google.com/s2/webfinger/? [email protected]       Experimental  Iden0ty:  WebFinger  
  24. h;p://www.x.com   h;p://slideshare.net/jcleblanc   User  Profile   h;p://www.google.com/profiles/nakedtechnologist   Portable

     Contacts   h;p://www-­‐opensocial.googleusercontent.com/api/ people/118167121283215553793/   Experimental  Iden0ty:  WebFinger  
  25. h;p://www.x.com   h;p://slideshare.net/jcleblanc   profileUrl   id   thumbnail  url

      urls   photos   Experimental  Iden0ty:  WebFinger   name        forma;ed        family  name        given  name        display  name    
  26. h;p://www.x.com   h;p://slideshare.net/jcleblanc   BrowserID  Source   <script  src="h;ps://browserid.org/include.js"  

    type="text/javascript"></script>     JQuery  Source   <script  src="h;p://code.jquery.com/jquery.min.js"   type="text/javascript"></script>   Experimental  Iden0ty:  BrowserID    
  27. h;p://www.x.com   h;p://slideshare.net/jcleblanc   Experimental  Iden0ty:  BrowserID     navigator.id.get(func0on(asserGon)

     {          if  (asserGon)  {                  $.ajax({                          url:  'h;ps://browserid.org/verify',                          type:  'POST',                          data:  'asserGon='+asserGon+'&audience=jcleblanc.com',                          success:  func0on(res)  {                                  console.log(res);                          }                  });   });  
  28. h;p://www.x.com   h;p://slideshare.net/jcleblanc   {          audience:

     "jcleblanc.com",          email:  "[email protected]",          expires:  1320081400987,          issuer:  "browserid.org",          status:  "okay"   }   Experimental  Iden0ty:  BrowserID  Results  
  29. h;p://www.x.com   h;p://slideshare.net/jcleblanc   The  FoundaGons  of  Human  IdenGty  

    Tribalism  and  Social  Grouping   Experimental  IdenGty  Methods   The  Big  Bag  of  Social  Iden0ty  Fail   What  We’re  Going  to  Cover  
  30. h;p://www.x.com   h;p://slideshare.net/jcleblanc       “My  privacy  concerns  are

     not  trite.   They  are  linked  to  my  actual   physical  safety”       -­‐-­‐Harriet  Jacobs  (Gizmodo)   Social  Iden0ty  Fail:  Personal  Safety   When  Social  Discovery  Impacts  Personal  Safety  
  31. h;p://www.x.com   h;p://slideshare.net/jcleblanc       “Path  Uploads  Your  EnGre

     iPhone   Contact  List  By  Default”       -­‐-­‐Mark  Hachman  (PCMag)   Social  Iden0ty  Fail:  Privacy  Concerns   When  Making  Things  Easy  Impairs  Privacy  
  32. h;p://www.x.com   h;p://slideshare.net/jcleblanc       “How  Target  Figured  Out

     A  Teen  Girl   Was  Pregnant  Before  Her  Father  Did”       -­‐-­‐Kashmir  Hill  (Forbes)   Social  Iden0ty  Fail:  The  Fine  Line   The  Fine  Line  Between  Insighbul  and  Creepy  
  33. h;p://www.x.com   h;p://slideshare.net/jcleblanc   IdenGty  is  more  than  just  a

     login     AuthenGcaGon  is  just  the  first  step     Find  the  tool  that:   – Has  the  raw  data  that  you  need   – Works  with  your  business   Iden0ty  Programming  Core  Concepts  
  34. hgp://slidesha.re/convergese_id   Thanks!  Any  Ques0ons?   Jonathan  LeBlanc   Developer

     Evangelist:  X.commerce   Twi;er:  @jcleblanc   E-­‐Mail:  [email protected]   Github:  github.com/jcleblanc