HiPPO and Flipism are no longer the only way to take decisions. In the Big Data / Data Science era one can dream of data-driven organization. If the data were "oil", Big Data technologies extract, transport, and store it, while Data Science methods provide the a way to "refine the crude oil". This presentation elaborates on the Ws (What, Why, When, Who and How) of Big Data and Data Science.

1. Big Data &
Data Science
W's
Emanuele Della Valle
@manudellavalle
Prof. @polimi & Founder @fluxedo_

2. W's
18/06/2018 @manudellavalle - http://emanueledellavalle.org 2

3. Why?
• In many organizations decisions are made by
"questionable" methodologies such as
– Highest Paid Person Opinion (HiPPO)
– Flipism (all decisions are made by flipping a coin)
18/06/2018 @manudellavalle - http://emanueledellavalle.org 3

4. Why?
Highest Paid Person Opinion (HiPPO)
18/06/2018 @manudellavalle - http://emanueledellavalle.org 4

5. Why?
Flipism (all decisions are made by flipping a coin)
18/06/2018 @manudellavalle - http://emanueledellavalle.org 5

6. Why?
• In many organizations decisions are made by the
"questionable" methodologies such as
– Highest Paid Person Opinion (HiPPO)
– Flipism (all decisions are made by flipping a coin)
• This could have been the right approach in the '70s …
– See the "Theory of Bounded Rationality" by Herbert Simons
18/06/2018 @manudellavalle - http://emanueledellavalle.org 6

7. Why?
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source http://www.azquotes.com/quote/139996 ]
7

8. Why?
• In many organizations decisions are made by the
"questionable" methodologies such as
– Highest Paid Person Opinion (HiPPO)
– Flipism (all decisions are made by flipping a coin)
• This could have been the right approach in the '70s …
– See the "Theory of Bounded Rationality" by Herbert Simons
• … but in the Big Data era one can dream of
data-driven organization
18/06/2018 @manudellavalle - http://emanueledellavalle.org 8

9. Why?
• Data-Driven Organization
18/06/2018 @manudellavalle - http://emanueledellavalle.org 9

10. Why?
Decisions no longer have to be made in the dark
or based on gut instinct; they can be based on
evidence, experiments and more accurate
forecasts.
-- McKinsey
18/06/2018 @manudellavalle - http://emanueledellavalle.org 10

11. Why?
• Data-driven organizations
– perform better
• The data shows where they can streamline their processes
– are operationally more predictable
• Data insights fuel current and future decision making
– are more profitable
• Constant improvements and better predictions help to
outsmart the competition and improve innovation.
18/06/2018 @manudellavalle - http://emanueledellavalle.org 11

12. Why?
• Moneyball: data + analysis to win games
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source: https://www.imdb.com/title/tt1210166/ ]
12

13. What's Big Data?
[source: IBM, 2012]
18/06/2018 @manudellavalle - http://emanueledellavalle.org 13

14. What's Big Data?
[source: IBM, 2012]
18/06/2018 @manudellavalle - http://emanueledellavalle.org 14

15. What's Big Data?
[source: IBM, 2012]
18/06/2018 @manudellavalle - http://emanueledellavalle.org 15

16. What's Big Data?
[source: IBM, 2012]
18/06/2018 @manudellavalle - http://emanueledellavalle.org 16

17. What's Big Data?
[source: IBM, 2012]
18/06/2018 @manudellavalle - http://emanueledellavalle.org 17

18. What's Big Data?
• Big Data is "crude oil" … that we have to
– Extract
– Transport in mega-tankers
– Ship through pipelines
– Store in massive silos
– …
18/06/2018 @manudellavalle - http://emanueledellavalle.org 18

19. What's Data Science?
• Data Science is "refining crude oil"
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source:http://allabtinstru.blogspot.com/2016/09/ProcessofRefiningCrudeOil.html]
19

20. What's Data Science?
• The Science [and Art] of…
– Discovering what we don’t know from data
– Obtaining predictive, actionable insight from data
– Creating Data Products that have business impact
now
– Communicating relevant business stories from data
– Building confidence in decisions that drive business
value
18/06/2018 @manudellavalle - http://emanueledellavalle.org 20

21. Who's a Data Scientist?
• Drew Conway, 2010
18/06/2018 @manudellavalle - http://emanueledellavalle.org 21

22. How?
• Statistics starts with data
• Two goals of analyzing data
– Descriptions: how nature associates responses to inputs
– Predictions: response for future input variables
[source: Statistical Modeling: The Two Cultures. Leo Breiman, 2001]
18/06/2018 @manudellavalle - http://emanueledellavalle.org
nature xy
independent
variable
response
variable
22

23. How?
[source: Marc Andrews, 2014]
Leverage more of the data being captured
18/06/2018 @manudellavalle - http://emanueledellavalle.org 23

24. How?
[source: Marc Andrews, 2014]
Leverage more of the data being captured
18/06/2018 @manudellavalle - http://emanueledellavalle.org 24

25. How?
[source: Marc Andrews, 2014]
Leverage more of the data being captured
18/06/2018 @manudellavalle - http://emanueledellavalle.org 25

26. How?
18/06/2018 @manudellavalle - http://emanueledellavalle.org
Reduce effort required to leverage data
[source: Marc Andrews, 2014]
26

27. How?
18/06/2018 @manudellavalle - http://emanueledellavalle.org
Reduce effort required to leverage data
[source: Marc Andrews, 2014]
27

28. What?
18/06/2018 @manudellavalle - http://emanueledellavalle.org
Reduce effort required to leverage data
[source: Marc Andrews, 2014]
28

29. How?
Data-driven exploration looking for correlation
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source: Marc Andrews, 2014]
29

30. How?
Data-driven exploration looking for correlation
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source: Marc Andrews, 2014]
30

31. Your butcher …
18/06/2018 @manudellavalle - http://emanueledellavalle.org 31

32. … at scale!
18/06/2018 @manudellavalle - http://emanueledellavalle.org 32

33. How?
Leverage data as it is captured
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source: Marc Andrews, 2014]
33

34. How?
Leverage data as it is captured
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source: Marc Andrews, 2014]
34

35. How?
Leverage data as it is captured
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source: Marc Andrews, 2014]
35

36. How?
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[sourcehttps://docs.microsoft.com/en-us/azure/machine-learning/team-data-science-process/]
36

37. How?
Overall picture by Gartner
18/06/2018 @manudellavalle - http://emanueledellavalle.org 37

38. Where?
18/06/2018 @manudellavalle - http://emanueledellavalle.org
[source https://www.ted.com/talks/anne_milgram_why_smart_statistics_are_the_key_to_fighting_crime ]
Improve public safety and
reduce violent crime
through data analytics
-41% murders | -27% crimes
38

39. Where?
18/06/2018 @manudellavalle - http://emanueledellavalle.org 39

40. Where?
18/06/2018 @manudellavalle - http://emanueledellavalle.org 40

41. What about cybersec?
18/06/2018 @manudellavalle - http://emanueledellavalle.org 41

42. Credits
• Big Data [sorry] & Data Science: What Does a Data Scientist Do?
Carlos Somohano, 2013
– https://www.slideshare.net/datasciencelondon/big-data-sorry-data-
science-what-does-a-data-scientist-do-world
• Becoming a data-driven organization The what, why and how.
SAS, 2018
– https://www.sas.com/en_us/whitepapers/becoming-data-driven-
organization-109150.html
• Never trust summary statistics alone; always visualize your data.
Alberto Cairo, 2016
– http://www.thefunctionalart.com/2016/08/download-datasaurus-
never-trust-summary.html
• 2017 Planning Guide for Data and Analytics. John Hagerty
(Gartner), 2016
– https://www.gartner.com/binaries/content/assets/events/keywords/
catalyst/catus8/2017_planning_guide_for_data_analytics.pdf
18/06/2018 @manudellavalle - http://emanueledellavalle.org 42

43. Thank you!
Any Question?
Emanuele Della Valle
@manudellavalle
Prof. @polimi & Founder @fluxedo_