Masterclass San Francisco: Data-driven analysis of social conversations using Natural Language Processing & the Brandwatch API

50 %
50 %
Information about Masterclass San Francisco: Data-driven analysis of social conversations...

Published on September 29, 2015

Author: brandwatchsocial

Source: slideshare.net

1. NLP  +  Brandwatch Analytics Deriving  insights  from  social  conversations using Natural   Language  Processing and  the  Brandwatch Analytics  API

2. What  we  will  cover  today What  do  we  want  to  answer?  (and  why) Our  approach  to  social  data Leveraging  the  Brandwatch API to  extract  data   Deriving  insight  from  personas   Identifying  key topics  of  conversation Segmenting  on  those  topics  to develop  personas Replicating  back  into  Brandwatch What  we’re  working  on  next

3. 3 | § Provides  an  in-­‐situ  portrait  based  on  exhibited  behavior  not   on  elicited  feedback § Highly  relevant  as  it  can  be  updated  in  near-­‐real  time § Enables  research  budget  to  be  focused  on  insights  rather   than  data  collection Social  intelligence  enables  new  ways  of   answering  traditional  business  questions   and  driving  data  driven  actions What  are  the  sort  of  questions  we  want  to  answer?   How  can  a  financial  services  company   reach  out  to  cyclists? How  can  we  get  small  business  owners   to  engage  with  their  cell  phone   provider  online? What  is  the  customer  journey  for  a   motorcycle  enthusiast?

4. 4 | There  are  seven  stages  to  the  analytical  process  of   developing  utilizing  personas  with  social  data Our  approach  to  working  with  social  data Extract Develop   the  dataset Linguistic   model Segment Analyze TrackQuery  data Prepare § Need  to  truly  understand  your  data  before   any  analysis § Iterative  query/dataset  development   through  virtual  ethnography § Use  the  Brandwatch API  to  extract  the  full   text  mentions   Model § Employ  Natural  Language   processing  to  model  how  people   talk § Use  either  qualitative  methods   or  clustering  algorithms  to   segment Understand § Through  visualization  and   analysis  we  can  understand   thoughts,  feelings  and   preferences § Replicate  back  into  Brandwatch as  sub-­‐categories  to  monitor  on   an  ongoing  basis

5. 5 | • Provide  the  basis  for  a  ‘corpus’  in  NLP  jargon  from  which   to  model • We  have  built  a  library  of  functions  using    python  to   retrieve  and  format  the  data • The  output  format  of  the  API  is  in  JSON  so  there  is  some   work  to  turn  it  into  a  table  we  can  read  and  use Extracted  BW  data  has  many  use  cases,  today  we   will  be  primarily  focused  on  full  text  mentions Leveraging  the  API  to  extract  the  dataset   Example  API  function: def get_mentions_query_URL(  startdate,enddate,project_id,   query_id,access_token,fullText): query_def =  "data/mentions”   end_date =  "endDate="  +  end_date +  "T00:00:00.000Z” start_date =  "startDate="  +  start_date +   "T00:00:00.000Z" request_URL="https://newapi.brandwatch.com/projects/"   +str(project_id)  +  "/"  +query_def if fullText ==  True: request_URL =  request_URL +  "/fulltext" request_URL =  request_URL +  "?"  +  "queryId="  +   str(query_id)  +  "&"  +  start_date +  "&"  +  end_date +   "&pageSize=5000"  +  "&access_token="  +  access_token return request_URL Read  more:  blog.tahzoo.com/tech-­‐thursday-­‐brandy-­‐py-­‐a-­‐python-­‐library-­‐for-­‐brandwatch/ Github:  https://github.com/BillmanH/brandy.py/

6. 6 | Linguistic  model  -­‐ Identifying  the  topics  in  a  conversation pumpkin sugar HEALTHY  LIVING PUMPKIN  SPICE CONVERSATIONS TEXT  ANALYSIS TOPIC   MODEL 1 Break  down  each  conversation  into  the  words  and   sentences  to  probabilistically  assess  each  word’s   relationship  with  each  other  word 2 Analyze  to  uncover  the  most  common  “topics”  of   conversation 3 Run  clustering  analysis  to  segment  on  topics 4 Iterate  on  topics  until  we  develop  a  solid  segmentation Four  steps  to  targeting  personas

7. 7 | ("pumpkin  spice  latte")   NOT ("vue pack"  OR "value  pack"  OR  "how  to  make"  OR   "win  free"  OR "latte  cake" OR  "black  friday"  OR "pack   of"  OR "My  TL  right  now  iOS7  Hump  Day  iOS7"   OR site:(twitter.com OR kdvr.com OR fox59.com  OR   news.google.com)) An  example:  who  discusses  Pumpkin  Spice  Lattes? Our  query… Excluded  because  of   irrelevant    recipes   Purposefully   broad   query  to  capture  full   range  of   conversations Exclude  Twitter  as  it   would  overwhelm  the   results  and  we  couldn’t   export  full  text  mentions

8. 8 | Do  it   Yourself Starbucks Nutrition Healthy   living Style Urban  living restaurants PS  recipes Amazing   treat Pumpkin   Spice   ingredients PS  Flavor Coffee  at   home people pumpkin grams squash fall city binary milk love food pumpkin home make spice fat healthy wear place victoire pumpkin time hari spice inch things latte calories recipe fashion food restaurants coffee day babe pie coffee life starbucks sugar food boots park options sugar good ingredients latte green time fall registers recipes style street time recipe back sugar flavor set thing psl data copycat color local pst spice week science flavored keurig feel drink saturated favorite wearing free visit cup great cancer seasonal count find coffee carbs soup dress art trading make home organic year price years today sodium paleo black event restaurant cream made found taste mountain world lattes pos version top restaurant september syrup work chemical food make The  topics  

9. 9 | DIY  Example: “I  get  annoyed  when  a recipe  calls  for  pumpkin  pie  spice.     It's  not  that  people  use  it  that  annoys  me,  it's  the  mere   existence  of  it  as  a  single  spice.    …  I  guess  I'm  just  a  purist   at  heart.    Since  I  haven't  seen  pumpkin  pie  spice  here  in   France  I  now  need  to  make  our  own  pumpkin  pie  spice   mixture,  and  then  figure  out  the  right  proportions  for  my   dreamboat  pumpkin  spice  latte.    Nothing  that  a  Google   search  won't  solve,  but  annoying  nonetheless.    And  don't   worry,  when  I  do  I'll  be  sure  to  share  it  with  you.    Maybe   you'll  even  get  some  rainbows.    Fingers  crossed.   An  example  of  how  this  analysis  works Treat/Reward  Routine  Example: “I  thought   splurging  on  a  venti pumpkin  spice  latte  would   make  me  feel  better  this  morning,  (or  maybe  even  the   three  cups  of  green  tea  with  lotsa honey  in  it!)  ...but  as   my  ears  pop,  my  nose  runs,  and  my  throat  feels  like   somebody  took  sandpaper  to  it  last  nite,  I  guess  it's  time   to  finally  suck  it  up  &  take  some  meds  ó¾Œ®ó¾ ‚  I  blame   you!  Rodney  Deal!!  Haha kidding  kidding  ;  ) Below  are  two  pieces  of  verbatim  content  that  we  used  in  our  model.  The  first  post  is  connected  with   the  DIY  (62%  relevant)  topic  and  the  second  with  Treat/Reward  (73%  relevant) 62% 73% DIY TREAT  /   REWARD PS  FLAVOR FALL  (SEASON) PSL  RECIPES HEALTHY   LIVING FILLER/   INFREQUENT   WORDS TOPICS:

10. 10 | • K-­‐means  highlights  clusters  of  conversations  based  on   the  topics  they  discuss • This  creates  a  segmentation  that  reflects  how  people   discuss  a  subject • Keys  in  on  the  pattern  of  topics  in  a  conversation   We  use  the  k  means  clustering  algorithm  to   segment  the  conversations  based  on  the  topics   in  order  to  create  the  personas Segmenting  on  the  topics  

11. 11 | Urban Living Fall   (season) Dessert Starbucks   drinks Pumpkin   flavor Treat  /     reward Pumpkin   (recipes) DIY Fall   (season) Treat  /   reward Pumpkin (flavor) Dessert StyleDesserts Treat /   Reward Being   Healthy Urban   Living Fall   (Season) Starbucks   Drinks Fall   (season) DIY Desserts Pumpkin spice   recipes Treat/   Reward LESS   IMPORTANT MORE   IMPORTANT Grouping  the  topics  that  are  core  to   each  segment  we  can  see  where   differences  break  down Mapping  topics  to  personas

12. 12 | Plotting  continuums  to  understand  the  personas Why  they  like  it What  it  stands  for NOVELTYNOSTALGIA GUILTY  PLEASURE DAILY  RITUAL OPPORTUNISTICTRADITIONAL PERENNIALSEASONAL OFTEN OCCASIONAL EXPECTED EARNED

13. 13 | What  we  found 22% PSL   PAMPERER A  pumpkin spice  latte  is  a  treat  to  be  savored   after  it’s  earned  or  after  a  tough  a  Monday   morning, “What  a  weekend.  Hello,  slow  Monday.  Oh  what's   that?  I  should  get  a  pumpkin  spice  latte?  Well,  if   you  insist...” 34% LATTE CHEMIST They make  their  own  lattes  in  the  comfort  of   their  own  home  or  tinker  with  the  official  version   “Here  is  an  awesome  home  version  of  Starbucks   Pumpkin  Spice  Latte.  Very  simple  to  make  and  alot cheaper… personally  I  like  it  better  because  you   control  the  amounts  of  ingredients  you  put  in  it   according  to  your  taste.” 38% FALL FANATIC Pumpkin  spice is  part  of  what  makes  fall  special   for  them,  a  pumpkin  spice  latte  is  one  part  of   their  fall  tradition “Pumpkin  Spice  Latte  at  Panera.  Oh  yeah,  I  need   one  of  those!    Bring  on  fall!    Looking  forward  to   bonfires  in  my  fire  pit  and  my  newly  refinished   fireplace.” 6% PUMPKIN TRADITIONALIST Loves  everything pumpkin  from  pumpkin  pie  to   lattes,  fall  is  just  an  excuse  to  get  their  fix  of   pumpkin “Are  you  ready  for  a  Pumpkin  Spice  Latte!?!?!    Or   how  about  a  Pumpkin  Bar????  Well   tomorrow  they   both  will  be  available!!!!”

14. 14 | Replicating  back  into  Brandwatch PSL  PAMPERER “morning  treat”  OR “Saved my  morning”  OR ((rough  OR bad   OR terrible*  OR awful  OR  stressful))  NEAR/4 (morning  OR day  OR  week)) LATTE CHEMIST ((my OR I  OR Mine  OR  “made  a”)  NEAR/2f (organic  OR make   OR recipe  OR mixture))  OR homemade  OR “the  perfect”  OR   ((coffee)  NEAR/3  (dessert  OR “sweet  tooth”) FALL FANATIC (I OR MY)  NEAR/3  (“love  fall”  OR “finally  here”  OR “the   season”  OR  autumn)  OR  ((making  OR  made  OR  bake  OR   baked)  NEAR/4f (cake  OR pie  OR pastry)) PUMPKIN TRADITIONALIST ((pumpkin) AND (candle  OR products OR  cake  OR pie))  OR “pumpkin  flavor”  OR ((“I  need  a”  OR “must  have”  OR “must   get”)  NEAR/3f  (latte)) We  conduct  a  careful  qualitative  analysis  of   persona  mentions  to  translate  the  topic   model  into  Brandwatch rules • Allows  us  to  visualize  and  track  in  Brandwatch • Create  each  persona  as  a  sub-­‐category • Creating  the  persona  rules are  iteratively  written Hypothetical  rules

15. 15 | Custom  geo-­‐mapping  for  DMA’s Persona  use  cases Typing  tools Scoring  conversation  relevance IDENTIFYING  TARGET   SEGMENTS Commentator DIY PS  Flavor Fall   (Season) Treat  /   Reward Focused  on  others Traditional

16. 16 Next  level  – what  we’re  working  on  now § Ability  to  use  the  model  to  tag  incoming  mentions   in  Brandwatch § Determining  demographic  characteristics  from   language § Utilizing  topics  to  predict  outcomes

17. 17 | Introduction  to  Tahzoo

18. 18 | What  we  do:  data  driven  customer  experiences Business  Oriented From  business  process  to  change  management,   Tahzoo  helps  enterprises  become  organizationally   ready  for  transformation Customer  Data  Centered Understanding  customers  from  their  own  point   of  view  is  the  foundation  for  successful   transformation,  Tahzoo  compiles,  curates  and   analyzes  all  the  data,  outside/in  and  inside  out Digital  Content  Inspired Customers  have  transformed  the  way  they   engage  with  companies  using  every  channel,   device  and  platform  available,  Tahzoo  addresses   engagement  through  content  marketing Technology  Enabled Architecting,  implementing  and  integrating  the   right  technical  solutions  determines  the   transformative  nature  of  the  experience,  this  is   Tahzoo’s  core  expertise  and  experience Frictionlessly  Delivered Both  the  customer  and  the  business  win  when  the   experience  is  delivered  as  efficiently  and  effectively   as  possible,  whether  that’s  delivered  in-­‐house,  as  a   managed  service  or  in  the  cloud  as  an  integrated   technology  and  marketing  operations  services   model

19. 19 | • Founded  in  2010 • Privately  Held • 300+  employees • Primary  Verticals • Financial  Services • Retail • Automotive/Manufacturing • Business  Services   A  fast  growing  CX  agency  focusing  on  digital   transformation  across  a  number  of  verticals Company  overview • Washington  DC  -­‐ HQ • Seattle,  WA • San  Francisco,  CA • Richmond,   VA • Milton  Keynes,  UK • Delft,  NL • Borlange,  SE • Novi  Sad,  RS Locations

20. 20 | Past  &  current  clients

21. 21 | Technology  partners

22. Thank  you Bill  Harding  – Data  Scientist billh@tahzoo.com Colin  Rogers  – Direction  of  Content  Strategy colinr@tahzoo.com

Add a comment

Related pages

Natural language processing - Wikipedia, the free encyclopedia

Natural language processing ... Using almost no information ... An explicit formalization of natural languages semantics without confusions with ...
Read more

Home | Mattermark

... and natural language processing to extract data from ... are using Mattermark to save time ... Products. Mattermark Pro; Mattermark API; Mattermark ...
Read more

EmTech Digital | MIT Technology Review

... San Francisco. ... MIT Technology Review EmTech Digital (formerly Digital Summit) ... Choose from six languages and in 13 regions worldwide.
Read more

DataSift | Human Data Intelligence

NUVI’s highly visual social media monitoring platform, combined with DataSift’s expertise in processing aggregate and anonymized Facebook topic data ...
Read more

Teradata - Global Leader in Data Warehousing, Big Data ...

Teradata is a leading provider of powerful, ... Data Driven Marketing, ... Corporate Social Responsibility;
Read more

Meltwater Group - Wikipedia, the free encyclopedia

Now headquartered in San Francisco, ... the billions of online conversations and ... that uses natural language processing technology to ...
Read more

Customer Intelligence Platform | Clarabridge

Make informed customer experience management ... Clarabridge built the industry-leading natural language processing and ... Sentiment Analysis; Social ...
Read more

Basis Technology, Experts in High Performance Multilingual ...

Linguistic analysis ... Airbnb’s Verified ID process uses Rosette Name Indexer to match names in multiple languages, ... Basis Technology provides ...
Read more