• Save
ANALYZING LARGE-SCALE USER DATA from Structure:Data 2012
Upcoming SlideShare
Loading in...5
×
 

ANALYZING LARGE-SCALE USER DATA from Structure:Data 2012

on

  • 473 views

Presentation from Aaron Kimball, WibiData

Presentation from Aaron Kimball, WibiData
#dataconf
More at http://event.gigaom.com/structuredata/

Statistics

Views

Total Views
473
Views on SlideShare
473
Embed Views
0

Actions

Likes
1
Downloads
0
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

ANALYZING LARGE-SCALE USER DATA from Structure:Data 2012 ANALYZING LARGE-SCALE USER DATA from Structure:Data 2012 Presentation Transcript

  • ANALYZING LARGE-SCALE USER DATA SPEAKER: Aaron Kimball CTO WibiDataFriday, July 27, 2012
  • Friday, July 27, 2012
  • Analyzing  Large-­‐Scale  User  Data with  Hadoop  and  HBase Aaron  Kimball  –  CTO WibiData,  Friday, July 27, 2012
  • We  can  now  collect   more  data  than  at   any  Dme  in  history.Friday, July 27, 2012
  • Yesterday’s  engineering   challenge:  FiJng  the   problem  into  the   hardware.Friday, July 27, 2012
  • Today’s  constrained   resource  is   understanding.Friday, July 27, 2012
  • How  do  we  best  apply   data …to  beMer  serving  our  Friday, July 27, 2012
  • The  best  products  are  user-­‐ • IntuiDve  UI • ConDnuously  learning – Guided  search – Smarter  recommenda1ons • More  effec1ve  serviceFriday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • What  are  we  building  Friday, July 27, 2012
  • Requirements 1.  Understand  the  user   populaDonFriday, July 27, 2012
  • Requirements 2.  Respond  to   users  in  real   DmeFriday, July 27, 2012
  • Requirements 3.  Support  graceful  data   evoluDonFriday, July 27, 2012
  • Large-­‐scale  data  science  is   • What  does  a  user  look  like? – What  data  is  available  about  the  user? – Which  features  are  important? – Which  features  are  correlated? • How  do  I  model  this  in  MapReduce? • How  do  I  serve  results  in  a  Dmely  Friday, July 27, 2012
  • Friday, July 27, 2012
  • Tools  of  the  trade • Store  all  data  about  a   user  in  one  place • Support  real-­‐Dme   get/put,  as  well  as   MapReduceFriday, July 27, 2012
  • Tools  of  the  trade • Use  complex  data   types  to  model   complex  data • Support  extended   data  models  over   DmeFriday, July 27, 2012
  • Tools  of  the  trade • Abstract  computaDonal   model  away  from   MapReduce • Support  computaDon   over  all  users…  or  one   user  at  a  DmeFriday, July 27, 2012
  • Friday, July 27, 2012
  • Friday, July 27, 2012
  • Friday, July 27, 2012
  • Friday, July 27, 2012
  • Friday, July 27, 2012
  • Friday, July 27, 2012
  • Friday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes Viewing/recording   historyFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes Viewing/recording   historyFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes          Libraries Device  and  User  Analysis Viewing/recording   history Personalized  offers   and   recommenda=onsFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes          Libraries Device  and  User  Analysis Viewing/recording   history Personalized  offers   and   recommenda=onsFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes          Libraries Device  and  User  Analysis Viewing/recording   history Personalized  offers   and   recommenda=ons Analysis  for   product   roadmapFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes          Libraries Device  and  User  Analysis Viewing/recording   history Personalized  offers   and   recommenda=ons Analysis  for   product   roadmapFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes          Libraries Device  and  User  Analysis Viewing/recording   history Personalized  offers   and   recommenda=ons Analysis  for   product   Tech  support   roadmap portalFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes          Libraries Device  and  User  Analysis Viewing/recording   history Personalized  offers   and   recommenda=ons Analysis  for   product   Tech  support   roadmap portalFriday, July 27, 2012
  •                                                      :  for  set-­‐top  boxes          Libraries Device  and  User  Analysis Viewing/recording   history Personalized  offers   and   recommenda=ons Improve Analysis  for   d  reports   product   Tech  support   for   roadmap portalFriday, July 27, 2012 adver=se
  • The  future • More  personalizaDon • AdapDve  UIs  (self  arranging   dashboards) • Targeted  content,  ads • More  effecDve  customer  serviceFriday, July 27, 2012
  • Conclusions • ApplicaDons  are  becoming   increasingly   user-­‐centric • Data  drives  this  capability,  but   harnessing  it  requires  a  new   distributed  architectureFriday, July 27, 2012
  • www.wibidata.com  /   Aaron  Kimball  –  aaron@wibidata.comFriday, July 27, 2012
  • Friday, July 27, 2012