Lars Hamberg
Predictive Analytics
in The Investment Industry
	
  
BIG DATA & ANALYTICS FOR BANKING SUMMIT
December 1 & 2, 2015 | New York
STRATEGIC SCOPE
OVERVIEW & OUTLOOK 	
  
QUESTION:
Has Big Data failed
	
  
QUESTION 1:
Has Big Data failed
…to deliver on its hype?
	
  
 
Big	
  data	
  defini'on?	
  
 
Big	
  data	
  defini'on?	
  
“too	
  large	
  OR	
  complex	
  	
  
for	
  tradi'onal	
  processing	
  methods”	
  
 
	
  	
  	
  general blur ?
buzzwords and distinction?
	
  
big	
  data	
  analy'cs	
  	
  
or	
  	
  
advanced	
  analy'cs	
  
?	
  
predic've	
  
analy'cs	
  	
  
or	
  	
  
big	
  data	
  
predic've	
  
analy'cs	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  
?	
  
	
  
machine	
  learning	
  ?	
  
	
  
	
  
 
	
  modern companies
want to know everything – what about banks?
	
  
structured	
  data	
  
	
  	
  
advanced	
  analy'cs	
  
	
  
	
  
staying	
  ahead	
  of	
  
the	
  connected	
  
clients	
  	
  
	
  	
  
predic've	
  
analy'cs	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  
	
  
machine	
  learning	
  	
  
	
  
	
  
QUESTION 2:
Why has Big Data failed
to deliver on its hype?
	
  
B I G D A T A – V I S I O N R U N N I N GA H E A D O F R E A L I T Y ?
BIG DATA HYPE CYCLE
Trigger
Enlightenment
Plateau of Productivity
Peak of Inflated Expectations
Trough of Disillusion
BIG DATA HYPE CYCLE
Trigger
Enlightenment
Plateau of Productivity
Peak of Inflated Expectations
Trough of Disillusion
BIG DATA HYPE CYCLE
WHY ARE WE HERE?
Trough of Disillusion
BIG DATA HYPE CYCLE
THE CHALLENGE OF
DEALING WITH THE REALITY
OF THE BIG UNSTRUCTURED
DATA STACK
Trough of Disillusion
In the big data stack
Unstructured
80-90%
2.5+	
  million	
  
words	
  in	
  english	
  	
  
QUESTION 3:
Will Big Data deliver on
its hype?
	
  
Yes!	
  
Why?
Confluence of trends
	
  
	
  
	
  
Confluence of trends
	
  1. Data explosion - 24m doubling rate
	
  
	
  
Confluence of trends
	
  1. Data explosion - 24m doubling rate
2. Smartphone user behavior – 150x/day
	
  
Confluence of trends
	
  1. Data explosion - 24m doubling rate
2. Smartphone user behavior – 150x/day
3. Breakthrough technologies: Machines
are learning to ”read and understand”
unstructured data on a large scale
Confluence of trends
	
  1. Data explosion - 24m doubling rate
2. Smartphone user behavior – 150x/day
3. Breakthrough technologies: Machines
are learning to ”read and understand”
unstructured data on a large scale (!)
Confluence of trends
	
  = the necessary conditions for
Big Data to deliver - and even
”over-deliver” - on its hype and
to transform many industries,
including the Investment
Industry
Breakthroughs	
  
Understanding	
  
the	
  meaning	
  	
  
of	
  words…	
  
Big	
  Data	
  Predic5ve	
  
Analy5cs	
  in	
  Produc5on	
  
and	
  in	
  Distribu5on	
  
Big	
  Data	
  Predic5ve	
  Analy5cs	
  
Magic	
  or	
  OSINT?	
  
 	
  	
  Magic	
  or	
  
OSINT?	
  
Big	
  Data	
  Predic5ve	
  
Analy5cs	
  in	
  Produc5on	
  
and	
  in	
  Distribu5on	
  
The mysteries of monetizing vast streams
of unstructured language data…
From data… …to dollars?
MUCH BETTER TEXT ANALYTICS
MUCH BETTER PREDICTIONS
MUCH BETTER SENTIMENTS
Magic	
  or	
  OSINT?	
  wha5smonitor.com	
  
Big	
  Data	
  Predic5ve	
  
Analy5cs	
  in	
  Produc5on	
  
and	
  in	
  Distribu5on	
  
I	
  am	
  special,	
  but	
  not	
  very	
  different…	
  	
  
Watson	
  Personality	
  Profile	
  (IBM)	
   Seman'c	
  Profile	
  (Gavagai)	
  
Use cases from financial industry
	
  
Distribution of investment products with
massive increases in profitability – Task &
Question
Common denominators for success
Common denominators for failure
Use cases from financial industry
	
  
Production of investment products –
alpha creation with big data predictive
analytics – Task & Question
Common denominators for success
Common denominators for failure
QUESTION 4:
Impact of Big Data
on your business?
	
  
Huge (transformative)
Imminent (surprising)
Winners & Losers (during the
shift – like in all major shifts)
	
  
Key take-aways:
Big Data has so far failed to deliver on its promise/hype – BUT
will over-deliver on its hype: Unsupervised, learning, scalable,
systems that understand the Big Data stack of language data,
enabling useful sentiment analysis and useful prediction
analytics on vast amounts of data
Predictive analytics, profiling, alias matching – use cases show
huge potential for disrupting most industries…Expect a
transformative shift in competitive landscape across most
industries – including banking
Banks – expect margin pressure, use case success stories in
analytics & opportunities for incumbents in and outside “non-
bank services” – uniquely positioned as distribution partners
Everybody can be a winner in this imminent shift – get involved
in advanced data analytics!
Lars Hamberg
Predictive Analytics
in The Investment Industry
	
  
BIG DATA & ANALYTICS FOR BANKING SUMMIT
December 1 & 2, 2015 | New York

Big data & analytics for banking new york lars hamberg

  • 1.
    Lars Hamberg Predictive Analytics inThe Investment Industry   BIG DATA & ANALYTICS FOR BANKING SUMMIT December 1 & 2, 2015 | New York
  • 2.
  • 3.
  • 4.
    QUESTION 1: Has BigData failed …to deliver on its hype?  
  • 5.
  • 6.
      Big  data  defini'on?   “too  large  OR  complex     for  tradi'onal  processing  methods”  
  • 7.
           general blur ? buzzwords and distinction?   big  data  analy'cs     or     advanced  analy'cs   ?   predic've   analy'cs     or     big  data   predic've   analy'cs                             ?     machine  learning  ?      
  • 8.
       modern companies wantto know everything – what about banks?   structured  data       advanced  analy'cs       staying  ahead  of   the  connected   clients         predic've   analy'cs                                   machine  learning        
  • 10.
    QUESTION 2: Why hasBig Data failed to deliver on its hype?  
  • 12.
    B I GD A T A – V I S I O N R U N N I N GA H E A D O F R E A L I T Y ?
  • 13.
    BIG DATA HYPECYCLE Trigger Enlightenment Plateau of Productivity Peak of Inflated Expectations Trough of Disillusion
  • 14.
    BIG DATA HYPECYCLE Trigger Enlightenment Plateau of Productivity Peak of Inflated Expectations Trough of Disillusion
  • 15.
    BIG DATA HYPECYCLE WHY ARE WE HERE? Trough of Disillusion
  • 16.
    BIG DATA HYPECYCLE THE CHALLENGE OF DEALING WITH THE REALITY OF THE BIG UNSTRUCTURED DATA STACK Trough of Disillusion
  • 17.
    In the bigdata stack Unstructured 80-90%
  • 18.
    2.5+  million   words  in  english    
  • 19.
    QUESTION 3: Will BigData deliver on its hype?  
  • 20.
  • 21.
  • 22.
  • 23.
    Confluence of trends  1. Data explosion - 24m doubling rate    
  • 24.
    Confluence of trends  1. Data explosion - 24m doubling rate 2. Smartphone user behavior – 150x/day  
  • 25.
    Confluence of trends  1. Data explosion - 24m doubling rate 2. Smartphone user behavior – 150x/day 3. Breakthrough technologies: Machines are learning to ”read and understand” unstructured data on a large scale
  • 26.
    Confluence of trends  1. Data explosion - 24m doubling rate 2. Smartphone user behavior – 150x/day 3. Breakthrough technologies: Machines are learning to ”read and understand” unstructured data on a large scale (!)
  • 27.
    Confluence of trends  = the necessary conditions for Big Data to deliver - and even ”over-deliver” - on its hype and to transform many industries, including the Investment Industry
  • 28.
  • 29.
    Understanding   the  meaning     of  words…  
  • 30.
    Big  Data  Predic5ve   Analy5cs  in  Produc5on   and  in  Distribu5on  
  • 31.
  • 32.
  • 34.
         Magic  or   OSINT?  
  • 35.
    Big  Data  Predic5ve   Analy5cs  in  Produc5on   and  in  Distribu5on  
  • 36.
    The mysteries ofmonetizing vast streams of unstructured language data… From data… …to dollars?
  • 37.
    MUCH BETTER TEXTANALYTICS MUCH BETTER PREDICTIONS MUCH BETTER SENTIMENTS
  • 38.
    Magic  or  OSINT?  wha5smonitor.com  
  • 39.
    Big  Data  Predic5ve   Analy5cs  in  Produc5on   and  in  Distribu5on  
  • 40.
    I  am  special,  but  not  very  different…    
  • 41.
    Watson  Personality  Profile  (IBM)   Seman'c  Profile  (Gavagai)  
  • 42.
    Use cases fromfinancial industry   Distribution of investment products with massive increases in profitability – Task & Question Common denominators for success Common denominators for failure
  • 43.
    Use cases fromfinancial industry   Production of investment products – alpha creation with big data predictive analytics – Task & Question Common denominators for success Common denominators for failure
  • 44.
    QUESTION 4: Impact ofBig Data on your business?  
  • 45.
    Huge (transformative) Imminent (surprising) Winners& Losers (during the shift – like in all major shifts)  
  • 46.
    Key take-aways: Big Datahas so far failed to deliver on its promise/hype – BUT will over-deliver on its hype: Unsupervised, learning, scalable, systems that understand the Big Data stack of language data, enabling useful sentiment analysis and useful prediction analytics on vast amounts of data Predictive analytics, profiling, alias matching – use cases show huge potential for disrupting most industries…Expect a transformative shift in competitive landscape across most industries – including banking Banks – expect margin pressure, use case success stories in analytics & opportunities for incumbents in and outside “non- bank services” – uniquely positioned as distribution partners Everybody can be a winner in this imminent shift – get involved in advanced data analytics!
  • 48.
    Lars Hamberg Predictive Analytics inThe Investment Industry   BIG DATA & ANALYTICS FOR BANKING SUMMIT December 1 & 2, 2015 | New York