WELCOME T O COMET 2017 - University of Texas at Austin · WELCOME T O TECHNICAL PAPER PRESENTATION...
Transcript of WELCOME T O COMET 2017 - University of Texas at Austin · WELCOME T O TECHNICAL PAPER PRESENTATION...
1 C O M E T 2 0 1 7
C O M E T 2 0 1 7 W E L C O M E T O
T E C H N I C A L P A P E R P R E S E N T A T I O N
2 C O M E T 2 0 1 7
B Y J A S O N C . D E N N I S O N
TRENDS IN DATA ANALYSIS
3 C O M E T 2 0 1 7
C R A I G S C H A U B C O N T R O L E N G I N E E R I N G M A N A G E R
A U S T I N E N E R G Y
A B O U T T H E A U T H O R
Craig Schaub manages a team of Engineers, Analysts and
Technicians who maintain Austin Energy’s real-time control systems
such as SCADA and ADMS.
He has more than 30 years’ experience in power system automation
as well as protection, telecommunications and metering.
J A S O N C . D E N N I S O N
M A N A G E R – L A B O R A T O R Y &
A N A L Y T I C A L S E R V I C E S
S D M Y E R S
A B O U T T H E A U T H O R
Jason C. Dennison manages the laboratory and analytical
services team at SDMyers, the world’s largest transformer
oil testing laboratory.
He has more than 14 years’ of wide experience in:
program management across multiple industries,
transformer and related equipment testing and
maintenance, software development, Internet of Things,
and technical education as an adjunct instructor. He has
his BS in Chemical Engineering with Polymer
Specialization from the University of Akron, is a Six Sigma
Black Belt, and maintains memberships with IEEE and
PMI.
BRIEF BIO
4 C O M E T 2 0 1 7
ABSTRACT SUMMARY
O V E R V I E W &
In th is d iscussion, we look at t rends in data analys is .
Data avai lab i l i ty is exploding, s tat ic and st reaming data
are col l id ing, and re l iab i l i ty profess ionals need to be
prepared to address the rapid ly expanding data analys is
needs wi th c lar i ty and conf idence.
L E A R N M O R E “… the possibilities of computers are very interesting – if they could be made to be more complicated by several
orders of magnitude. If they had millions of times as many elements they could make judgments.”
– Richard Feynman, “There’s Plenty of Room at the Bottom” 1959
5 C O M E T 2 0 1 7
TREND: DATA CHAOS
Honeywell: In 2017, we est imate that only 1% of IoT data is
currently used to add value.
By 2025 IoT devices wi l l be generat ing 136 terabytes of
data annual ly.
A lso, the market for IoT wi l l be $11 Tri l l ion USD .
6 C O M E T 2 0 1 7
TREND: STATIC & STREAMING DATA
IDC Data Age 2025: By 2025 the s ize of the g lobal
datasphere wi l l reach 163 ZETTABYTES .
IoT / Real-Time Data wi l l compr ise near ly 30% of a l l
g lobal data in that t ime.
McKinsey Global Inst i tu te: IoT data is only being used
for detect ing anomal ies…not for opt imizat ion and
predic t ion, which prov ide the greatest va lue.
Today – we use a fract ion of a fract ion of avai lab le
data AND our decision-making hasn’t caught up with
the amount of data we do have.
163 Zettabytes represents >650% growth in the next 7 years
7 C O M E T 2 0 1 7
DATA WAREHOUSE DATA LAKE
AGGREGATE
DATA SOURCES
STRUCTURE
REPORTING
COMPILE RAW
DATA SOURCES
STRUCTURE
REPORTING
?
?
GOOD: Structured data for wide usage by users of varying expertise BAD: Structuring the data takes definition, time, and developer resources
GOOD: No structure make acquisition simple, fast BAD: Structuring the data now the responsibility of the user/requestor
TREND: BLENDED DATA ARCHITECTURE
8 C O M E T 2 0 1 7
TREND: DATA EXTRACTION and ABSTRACTION
Data storage is increasingly HYBRID – Data
warehouses, Data lakes, SQL/MySQL/Oracle/Mongo
Databases, Operat ional f i les (Excel , Access) , Cloud
Serv ices
Data ext ract ion is cruc ia l in al igning data f rom
disparate sources, so that analys is is truthful .
Good computer scientists and analysts wi l l be the
heroes of the Informat ion Age. Good designers wi l l be
the unsung heroes of the Informat ion Age.
Structured Data
Warehouses, Data Lakes
Org/Ops Functional Databases
(SCADA, M/O, Excel)
Could Services (Salesforce,
ConstantContact)
Visualization (Tableau, Qlik, PowerBI) Statics (R) Machine Learning AI Custom App
D ATA E X T R A C T I O N ( Q U E R I E S ,
S TO R E D P R O C E D U R E S , D L L S )
9 C O M E T 2 0 1 7
EXAMPLE: STATIC VS STREAMING
With Single Point Data, personnel are tasked wi th
decis ion-making – lack ing facts, nature is to assume the
worst or ignore the possibi l i ty of an issue
With h igher quant i ty data (assumed to be of h igh
in tegr i ty) , t rends are more apparent , and the t iming on
the P-F curve can be orders of magni tude smal ler
MONTHS
MINUTES
10 C O M E T 2 0 1 7
TREND: DATA SCIENCE GROWTH
The only non-heal thcare job growth area in the top 5 is
Mathemat ica l Science Occupat ions, re lated to analyt ics
and stat is t ics
In the TOP 10, a s tagger ing 8 of 10 are heal thcare
re lated; the other entry is o i l /gas/min ing serv ice uni t
operators
11 C O M E T 2 0 1 7
TREND: IT GROWTH
Cyber secur i ty
NERC – Interpretat ion & Implementat ion
IT best pract ices
Regulat ion
“Once you are inside [the system], the assumption is that you are supposed to be there.”
– Richard Bejtlich, Chief Security Strategst @ FireEye
“Every budget is an IT budget.”
– Erwin Verstraelen, CIO @ AVEVE
12 C O M E T 2 0 1 7
TREND: DECISION RIGHTS MANAGEMENT
Decis ion r ights management (DRM) isn ’ t brand
new, though the data explos ion that dr ives DRM
value is .
The key: Clear ly ident i fy who is / who needs to be
RESPONSIBLE for a decis ion/ task, and
EMPOWER them to make decis ions by way of due
process and c lear co l laborat ion path.
Resp
onsibility
VP
Director
Manager
Supervisor
Subject M
atter Expert
Task DescriptionTask 1 I I C RA CTask 2 R A C CTask 3 R A CTask 4 I ITask 5 R CTask 6 I A RA C
R Responsible for ensuring completionA Accountable for task or providing informationC Consulted for the task and/or informationI Informed of the task
13 C O M E T 2 0 1 7
THE HORIZON
Machine learn ing & AI
Natura l language processing
SCADA … in the c loud!
14 C O M E T 2 0 1 7
SUMMARY
The dawn of the INFORMATION AGE is here and in i t ’s
in fancy; the natura l order is DISORDER .
We’re now able to have fa i lure data near ly
INSTANTANEOUSLY , and whi le h igh value, is creat ing
chaos.
Manpower, Machines, Measurements, Methods must be
robust to address new information quick ly and
EMPOWER exper ts to take decisive, intel l igent act ion
Star t wi th the end in mind, and GET STARTED - 80%
effect iveness NOW is more valuable than 100% LATER
15 C O M E T 2 0 1 7
P A P E R Q & A C O M E T | C O L L A B O R A T E | S H A R E
K N O W L E D G E | P R O G R E S S
16 C O M E T 2 0 1 7
3 3 0 . 2 8 9 . 3 4 2 1 j a s o n . d e n n i s o n @ s d m y e r s . c o m linkedin.com/in/ jasoncdennison
JASON C. DENNISON
MANAGER OF LABORATORY & ANALYTICAL SERVICES
GET IN TOUCH