Trends in Big Data 
René Kuipers 
Principal Consultant Big Data & Analytics 
@rjlkuipers
Overview 
• Wat is it (or not) ? 
• What to do with it ? 
• How to deal with it ? 
• Impact on IT?
Big Data 
Wat is it (or not)
Big data: what is it (or not)
Big data: what is it (or not) 
• Social data. 
• Sensor data. 
• unstructured data. 
• A hype 
• Hadoop 
• Real-time 
• ……………… 
• The four V's: Volume, Variety, Velocity, Value
Big data: what is it (or not)
Wat exactly is the trend ? 
• “a lot of data” isn’t new: 
• Governments 
• Stock exchange. 
• Real-time isn’t new: 
– Railway management 
• Level of detail increases 
• Data is no longer in possession 
• Open Data intiatives.
Example 1: 
on-board sensor data 
• America’s Cup 2013 
• On-board sensors 
• Off-board sensors 
• 30.000 dp/s 
• Oracle Team USA won 9-8 after 8-1 down, because of ‘insight’.
Example 2: 
medical 
• personalized healthcare 
• n=1 treatment. 
• Bring down cohort-size by 
means of more personal 
data.
Example 3: 
scientific 
• CERN (LHC) 
• 120.000 sensors 
• 4 GB data/sec
Example 4: 
online ads 
• Real-time profiling 
• Context search 
• Display ad 
• Billing
Big Data Social 
Cloud Mobile 
Een unity of 4
The value of data
Value of data 
• The higher up in an 
organization, the lower the 
value of an individual 
datapoint. 
• The lower (more operational) 
in an organization, the higher 
the value of an individual 
datapoint. 
• Big data offers more (real-time) 
insight on 
operational level than on 
strategic level.
• Big Data is mainly a CX ‘thing’. 
• 360-degree view of 
• Patient 
• Customer 
• Supplier 
• … 
• …
Implications 
Organization Technical
Implications 
Organization 
– Outward. 
– Openness, transparency 
– Extreme short communication (outward) 
– Customer focus, short reactiontimes.
Customerfocus 
#klm #fail
Implicaties 
Organization Technical
How to embed Big Data into traditional DWH/BI ?
How to embed Big Data into traditional DWH/BI ?
How to embed Big Data into traditional DWH/BI ?
DWH and BI 
changes
In-house 
data 
External 
data
Datawarehouse, Big Data and Business Intelligence 
– Storage of Big Data: Hadoop 
– Storage of ‘traditional’ data: RDBMS 
– necessity: transparant access. 
– necessity: high-end usertools: Self Service BI. 
– necessity: fast back-end.
Future
More data More 
questions 
Faster 
answers
In-memory
Conclusions 
– Big Data 
– Large Volume (Volume) 
– Fast (Velocity) 
– In different shapes and sizes (Variety) 
– Huge information potential (Value) 
– Data grows exponentially 
– Data is extern 
– Data must lead to faster insights 
– Fundamentallly datawarehouse redesign 
– In-memory 
– Demands organizational changes
Presentation Big Data

Presentation Big Data

  • 2.
    Trends in BigData René Kuipers Principal Consultant Big Data & Analytics @rjlkuipers
  • 3.
    Overview • Watis it (or not) ? • What to do with it ? • How to deal with it ? • Impact on IT?
  • 4.
    Big Data Watis it (or not)
  • 5.
    Big data: whatis it (or not)
  • 6.
    Big data: whatis it (or not) • Social data. • Sensor data. • unstructured data. • A hype • Hadoop • Real-time • ……………… • The four V's: Volume, Variety, Velocity, Value
  • 7.
    Big data: whatis it (or not)
  • 9.
    Wat exactly isthe trend ? • “a lot of data” isn’t new: • Governments • Stock exchange. • Real-time isn’t new: – Railway management • Level of detail increases • Data is no longer in possession • Open Data intiatives.
  • 11.
    Example 1: on-boardsensor data • America’s Cup 2013 • On-board sensors • Off-board sensors • 30.000 dp/s • Oracle Team USA won 9-8 after 8-1 down, because of ‘insight’.
  • 12.
    Example 2: medical • personalized healthcare • n=1 treatment. • Bring down cohort-size by means of more personal data.
  • 13.
    Example 3: scientific • CERN (LHC) • 120.000 sensors • 4 GB data/sec
  • 14.
    Example 4: onlineads • Real-time profiling • Context search • Display ad • Billing
  • 15.
    Big Data Social Cloud Mobile Een unity of 4
  • 16.
  • 17.
    Value of data • The higher up in an organization, the lower the value of an individual datapoint. • The lower (more operational) in an organization, the higher the value of an individual datapoint. • Big data offers more (real-time) insight on operational level than on strategic level.
  • 18.
    • Big Datais mainly a CX ‘thing’. • 360-degree view of • Patient • Customer • Supplier • … • …
  • 19.
  • 20.
    Implications Organization –Outward. – Openness, transparency – Extreme short communication (outward) – Customer focus, short reactiontimes.
  • 21.
  • 22.
  • 23.
    How to embedBig Data into traditional DWH/BI ?
  • 24.
    How to embedBig Data into traditional DWH/BI ?
  • 25.
    How to embedBig Data into traditional DWH/BI ?
  • 26.
    DWH and BI changes
  • 27.
  • 28.
    Datawarehouse, Big Dataand Business Intelligence – Storage of Big Data: Hadoop – Storage of ‘traditional’ data: RDBMS – necessity: transparant access. – necessity: high-end usertools: Self Service BI. – necessity: fast back-end.
  • 29.
  • 30.
    More data More questions Faster answers
  • 31.
  • 32.
    Conclusions – BigData – Large Volume (Volume) – Fast (Velocity) – In different shapes and sizes (Variety) – Huge information potential (Value) – Data grows exponentially – Data is extern – Data must lead to faster insights – Fundamentallly datawarehouse redesign – In-memory – Demands organizational changes

Editor's Notes

  • #9 Opmerkelijk: dit gaat over aantallen, over groei.
  • #10 Detailniveau wordt groter: “Logistiek voorbeeld: waar is de vrachtwagen nu ?, versus aan het eind van de dag een bon afleveren.”
  • #12 72.000.000 datapunten per race.
  • #17 De data-waarde-keten
  • #25 Hoe Big Data aan te haken hangt van de toepassing af: real-time actie (AddJuggler, fraude detectie van pintransacties vereisen een andere implementatie) Hier Scania voorbeeld
  • #26 Hoe Big Data aan te haken hangt van de toepassing af: real-time actie (AddJuggler, fraude detectie van pintransacties vereisen een andere implementatie)
  • #27 Hoe Big Data aan te haken hangt van de toepassing af: real-time actie (AddJuggler, fraude detectie van pintransacties vereisen een andere implementatie)