p 02 ta_in_uw_transformation_2017_06_13_v5
TRANSCRIPT
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
Life Insurance
Transformation (Using Text Analytics)
Vishwa Kolla Head of Advanced Analytics,
John Hancock Insurance
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
2
Advanced
Analytics CoE,
Maturity Model
Customer Analytics
(entire value chain)
Machine Learning
Scoring Engine
Optimization
Simulations
Forecasting & Time
Series
• 16+ Years
• John Hancock Insurance
• Deloitte Consulting (Industries – Insurance,
Retail, Financial, Technology, Telecom,
Healthcare, Data)
• IBM
• Sun Microsystems
Business Analytical (Math, Stats)
Technical (Programming)
Expertise
Experience
Vishwa Kolla Head of Advanced Analytics
John Hancock Insurance, Boston
MBA Carnegie Mellon University
MS University of Denver
BS BITS Pilani, India
4
Where How So
What?
Unigram Unigram analysis is a
stat. Be cautious – can
lead to a wrong
conclusion if close
attention is not paid
6
Where How So
What?
Unigram Trigram and above Bigram
A tri-gram analysis
seems more like an
over-kill for this
particular problem.
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
7
Prospect Acquire Nurture
Always think simple in
3s.
When thinking about
adding value, we can
broadly think of 3 areas.
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
8
Prospect Acquire Nurture
• Document Clustering /
Classification
• Word Cloud
• Concept Extraction
• NLP
Social Listening When looking to see
who to target, social
listening seems to be a
good direction
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
9
Prospect Acquire Nurture
• Info Retrieval / OCR
• Document Clustering /
Classification
• Concept Extraction
• NLP
• Web Mining
Information Organization When looking to
improve operations,
information extraction
and structuring is more
important
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
10
Prospect Acquire Nurture
• Document Clustering /
Classification
• Concept Extraction
• NLP
• Web Mining
Information Understanding When looking to nurture
existing relationships,
text analytics of
interactions help a ton
11
Where How So
What?
Benchmark
#PASANDIEGO #ODSC+TDWI+MLSUBMIT+DS AUSTIN
An insight is more
powerful when we
benchmark
12
#PASANDIEGO
#DATA SCIENCE CON AUSTIN
#ODSCEAST
#TDWI
#Machine Learning Summit
Where How So
What?
Benchmark 1 vs. all and
1 vs. other
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
13
https://www.youtube.com/watch?feature=player_embedded&v=XswX1TSQwfQ
So many stories. IoT led
disruption. Shared
Value. Business Model
Disruption. Breathe new
life into Life Insurance
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
14
• Where should I talk?
• What topics should I talk about?
• Who should I talk to? Word
Cloud
Classification Alignment
Inventory EDA Relevance
Monitor
Influencer
Pool
Relevance Partner
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
15
What are
people talking
about?
Why do we have a
spike in positive
tweets?
Has our NPS
increased since
we launched ? NPS
18.52%
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
• How do I extract
value from Dark data?
• What APSes are
similar?
• What is the insight?
• Is there a misrep?
• What does my customer want?
• Is my customer a
Net Promoter?
• How can I nudge?
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
17
• How do I extract
value from Dark data?
• What APSes are
similar?
• Is there a misrep?
• What does my customer want?
• Is my customer a
Net Promoter?
• How can I nudge?
Shiny R
App
Web scrapping
Sentiment Analysis
Word clouds
Topic Modelling
Word Frequency
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
19
Increased CSAT
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
21
Increased NPS
Reduced Software Spending
Reduced Services Spending
Always quantify value
Global
Big
Data
Confe
rence,
Santa
Clara
March
7-9
22
Means to an End
Can be very distracting
Automate Mundane
Tasks into Packages
Text Analytics = more
than R&D
Use tools to your
advantage