Writing
up quick-hits I'm going to post a month from now. When I read them again,
pre-post, I'll probably forget I wrote them. "I say that?"
"Record
Store Day" approacheth. Remember when it was no big deal to hang out in
record stores? Now it's the mark of an audiophile supergeek
Just
saw reference to "David Cameroon, prime minister of the United
Kingdom." Like Obama, he's the first half-African to govern that country
"The
Data-Visualization Revolution" (http://ow.ly/vSYzB )
JK--Displays Brazil's cool DataViva visualizations of open public data.
"AllJoyn:
A Common Language 4 IoT" (http://ow.ly/vSY3V ) JK--Developed by
AllSeen Alliance. Open source. Enable hetero IoT sol'n interop
"What
do Biz Intelligence Consultants Mean by 'Solutions'?" (http://ow.ly/vSXPr )
JK--#Forrester
@bevelson nicely parses distinctions
I
never fail to be amazed by IT industry's ability to publish article after
article saying substantially the same things, slightly reworded
An
optimist is an optometrist who only fits you with rose-colored prescriptions.
PBS "Nature." Naturalist
attempts to "become one" with a mule deer family. Daughter & I
agree that's slightly unsettling.
CNN
seems to have elevated the missing Malaysian airliner story to full 9/11
status. 24x7 saturation coverage weeks after the incident.
Today's
been "pulling my hair out of my head" busy. And, as you can see, I
don't have much to spare.
A8:
#bigdatamgmt
Recognize that data gov/mgt policies may differ gr8ly among domains. Official
system of records data gov is most stringent
A8:
#bigdatamgmt
Establish top-down lifecycle data governance/stewardship policies/practices
that are consistent across all domains.
A8:
#bigdatamgmt
Assign biz-level responsibility for data stewardship of domains (customer,
finance, etc.). They "own" it across lifecycle
A8:
#bigdatamgmt
Inventory your data top to bottom. You can't manage what you don't know exists.
A7:Apply
role based AA. Audit & monitor usage & data flows. Anonymize &
encrypt sensitive data as much as possible #bigdatamgmt
Retweeted
by jameskobielus
A7
Should be able to classify which files in a Big Data platform contain sensitive
data and to know where that data resides #bigdatamgmt
Retweeted
by jameskobielus
A7:
#bigdatamgmt
Mitigating risk of NOT being able to comply with key data regs/mandates
involves CIO-level planning & ongoing monitoring
A7:
#bigdatamgmt
Mitigating risk of your business NOT finding non-obvious insights in your #bigdata
involves staffing up on data scientists
A7:
#bigdatamgmt
Mitigating technical risks involves 24x7 high availability, security, &
performance on all your #bigdata
platforms
A7:
#bigdatamgmt
Mitigating biz risk starts with proverbial "single version of truth"
with robust #bigdata
governance across all domains
A6:
#bigdatamgmt
Rolling out new data-driven functionalty faster & more securely also
requires strong data governance tooling.
A6:
#bigdatamgmt
Test data retrieved from shared platform should be masked using standard tools
such as #IBM
Optim
A6:
#bigdatamgmt
Agility = reuse existing data, analytics, & plats. App devs should reuse
test data from fast, secure, shared #bigdata plat
@RyMontano #bigdatamgmt
A5: Adoption of Spark & other #bigdata plats
not coming at #Hadoop
expense. Complementary. New niches 4 new plats
A5:
#bigdatamgmt
Chief #Hadoop
niches include 1) staging, 2) sandboxing, & 3) archiving. All involve
moving data in&out of Hadoop.
A5:
#bigdatamgmt
Ambiguous question has a 3rd possible interpretation: "Hadoop niche plat
within data archs?" Answer is "yes, increasingly."
A5:
#bigdatamgmt
This is ambiguous. Q could mean "#Hadoop becoming
source platform?" or "people abandoning Hadoop?" Answer to both
is no
A4.
Real-time analytics on real-time data means stream processing on data in motion
#bigdatamgmt
Retweeted
by jameskobielus
A4:
Integration of DW & Hadoop is possible as well to get the best of both #BigDataMgmt
http://ow.ly/vPPkS
Retweeted
by jameskobielus
A4.
Exploratory analytics on schema less multi-structured data will likely choose
Hadoop #bigdatamgmt
Retweeted
by jameskobielus
#bigdatamgmt
A4: People know where EDWs excel, but still have the open q in their minds
"Will Hadoop be as good/cheaper?"
Retweeted
by jameskobielus
A4:
Most CIO's don't believe Hadoop is going 2 replace EDW, each will have its
place #BigDataMgmt
http://ow.ly/vPPqN
Retweeted
by jameskobielus
A4:
#bigdatamgmt
Consider "where data should go" & "where workloads should
go" issues jointly. In-db analytics says wklds should go to data
A4:
#bigdatamgmt
Data doesn't live in just one platform throughout its lifecycle. It
"should go" to any/all them, depending on workloads
A4:
#bigdatamgmt
Tiered arch o fit-to-purpose data plats: source (OLTP), staging (Hadoop,
NoSQL), EDW (RDBMS), access (in-mem, colmnr), etc
A3:
Some industries keep data 50-99 years, eventually value declines. Don't have to
store it all on expensive platform. #BigDataMgmt
Retweeted
by jameskobielus
A3:Think
from C's PoV. Can u find valuable use cases utilizing "old" data? If
not move it 2 archs 4 #DataScientist
to explore #bigdatamgmt
Retweeted
by jameskobielus
A3:
#bigdatamgmt
Store older/less-used data on cheaper/bulk/offline storage & new/more-used
data on fast, online storage platforms.
A3:
#bigdatamgmt
Store data only as long as a) mandated by extl/internal requirements and b)
provides biz value outweighing carrying cost
A3:
#bigdatamgmt
Store the data sets needed for current/anticipated operational reporting,
decision support, perf mgt, mining, compliance
A2:
Store until decisions made from #bigdata are
very minimally marginal and have a higher cost than benefit. #bigdatamgmt
Retweeted
by jameskobielus
70%
of enterprise data is debris, 50-75% of OLTP data is dormant, why keep it? #BigDataMgmt http://ow.ly/vPOVm
A2:
#BigData
storage is essential for data journalists though. Records and archives makes so
many of those viral infographics #bigdatamgmt
Retweeted
by jameskobielus
A2:
#bigdatamgmt
Rarely true that "don't know when might need" older data. Specific
future scenarios (eg compliance rptg) might be clear
A2:
#bigdatamgmt
Priority data reqs--present & anticipated--must drive retention. If don't
foresee scenario where old data needed, purge
A2:
#bigdatamgmt
Storing "everything" only makes sense if data tiny or if storage
free/infinite & retrieval magically fast & precise
A2:
#bigdatamgmt
Hoarding is never wise, in any sphere of human activity, data or otherwise.
Curating, archiving, & purging always essential
A1:operational
data has a sell-by date for its intended purpose,for analytics,could be
foreverAnalytical data is "used" #bigdatamgmt
Retweeted
by jameskobielus
A1:
#bigdatamgmt
"Typicality" of #bigdata
lifecycle varies by platforms. #Hadoop data
mgmt lifecycle differs from OLTP, RDBMS, EDW, etc
A1:
#bigdatamgmt
Lifecycle of unstructured #bigdata often
quite different from structured data: discovery, profiling, analysis, use, etc
A1:
#bigdatamgmt
"Bigness" of #bigdata
intensifies lifecycle challenges: scaling, optimization, protection,
exploration, etc.
A1:
#bigdatamgmt
#BigData
lifecyle is data management lifecycle: originate, acquire, integrate, process,
govern, analyze, deliver, use.
@IBMbigdata Happy to snapping my
fingers to the keyboard on today's #bigdatamgmt
chat. Lifecycle lifecycle!
RT
@IBMbigdata: 22 mins to #bigdatamgmt
chat w/ @mikeferguson1, @BTRG_MikeMartin and @jameskobielus. Join us! http://bit.ly/1eJkPKJ
RT
@kexpplaylist #kexp Wash Yr
Hands by Falcons Of Fine Dining from Demo JK--2014. I thought this was a lost
Devo track. Derivative but good
You'll
see blogs from me weekly while I'm away. I've pre-written 7 of them. To be
posted by the various editors.
Next
quick-hit will be Monday May 19. Vacation & business trip between then and
now. I work tomorrow, but wrapping stuff pre-vacation.
Context accumulation? Grounding
cognitive confidence in the probabilistic fabric of the real world (http://ow.ly/vQv5E )
Wed #IBM q-h
RT
@kexpplaylist #kexp Making Plans
4 Nigel by XTC from Waxworks: Some Singles 77-82 JK--Plans 4 Nigel 2 execute or
is plan 2 execute Nigel?
People
promise stuff w/reckless abandon. Always good 2 assign them tiny token task 2
gauge their sincerity. See if can B relied on 4 largers
One
of life's greatest challenges is to get people to sit still so you can finish a
sentence. That includes simple sentences. Even fragments
@Pini_Yakuel What
"approach"? It's just a metaphor.
"Predictive
analysis to identify who will revisit your site in next few days" (http://ow.ly/vPe5f )
JK--Article specifies key data points
Well,
at least they didn't call the vulnerability "Aortic Aneurysm."
"Using
#IBM #Watson
Foundations to read emotions" (http://ow.ly/vOV6I )
JK--Great piece by Joe Clabby on facial imaging.
Internet
o Things? Th binocular vision & opposable thumb o cognitiv cmptg p1: http://ow.ly/vO0Qf
p2: http://ow.ly/vO0Rt
Tues #IBM qh
Orwell's
"Big Brother." Mine's never been totalitarian 2 me, nor I to my
youngers. But I suppose there's a little bit of Stalin in all bros
"Bluetooth
Wants Its Bite o #BigData"
(http://ow.ly/vNYCK )
JK--For IoT wearable wirelss device connect 2 feed/consume data to/from cloud
RT
@kexpplaylist #kexp My Girls by
Tears For Fears from My Girls JK--2014. Animal Collective covr by moody indie
duo best known for 80s hits
"Music
analytics is helping music industry see into future" (http://ow.ly/vNXhw )
JK--But that's just sensing applause, not quality.
"Data:
The Modern Tribal Bonfire" (http://ow.ly/vNW1m ) JK--If
McLuhan were still around, I'd yank him into this discussion, a la Woody
Thinking
of starting my next blog by tracing the origin of #bigdata back
to caveman days. Not.
I've
never bought into this "the PC is dead" jazz. A laptop is just a PC
you carry around with you. I'm typing this on one now. It's alive.
"Building
a #Hadoop DW:
Dr. Ralph Kimball Answers Your Questions (http://ow.ly/vNRjp )
JK--Excellent practical detail in his responses.
Join
me et al on #BigDataMgmt
chat "Managing the lifecycle of #bigdata"
tomorrow 12pm ET http://bit.ly/1eqwwpp
#Watson won the
Masters? We approve!
A8
A good Chief Customer Officer can tie it all together across silos + create
unified, internal understanding. #CXO
Retweeted
by jameskobielus
A8:
#cxo Business
should do CXO in a closed-loop: improving satis metrics which drives
recommendations which drives cust acquisitions
A8:
#cxo Customer
Influence, customer word of mouth, & customer acquisition: a happy customer
tells the world.
A7:
Give them access to the same data (for starters), and then build
cross-functional team w/authority to review/react to that data #CXO
Retweeted
by jameskobielus
A7:
#cxo #Custserv
teams must collab with #quality
assurance teams to define key quality KPIs highlighted in operational analytics
A7:
#cxo Orgs should
converge #quality
& #custserv
roles. Quality assurance must be in core skillset of all customer-facing roles
A6:
#cxo If your
quality-relevant #bigdata
is primarily in an exploratory sandbox that doesn't drive operational
decisions, it's useless
A6:
#cxo To leverage
#bigdata for
improving service quality, need it to feed useful analytics (reports, KPIs,
etc.) for decision makers
A6:
#cxo If you fail
to use your #bigdata
to find service problems before the customer churns, that's a huge problem.
A6:
#cxo By itself, #bigdata
doesn't help or hinder service quality. Success (or lack thereof) in assure
qual requires U mine that data
A5:
#cxo Fixing gaps
in service quality may require closed-loop remediation/data gathering with A/B
testing.
A5:
#cxo Gaps in
service quality may be complex and involve many variables over time. Do
historical & predictive stat analysis.
A5:
#cxo If you
provide products/services that generate Internet of Things data, and customers
grant access, may show service degradation
A5:
#cxo If you're
doing e-commerce, web analytics data (eg., browser & cart abandonments) may
signal low-quality experience
A5:
#cxo VoC data is
a perfect source of intelligence to identify service quality gaps.
A4:
#cxo Don't be in
thrall to your own marketing message. Level of cynicism that customers express
shows how hollow they consider it 2 B
A4:
#cxo Ongoing
Voice of the Customer (VoC) programs are essential for tuning your listening to
what they're really saying.
A4:
#cxo Customers
will tell you whether the value they receive matches what your marketing
boasts. And they'll tell you bluntly.
A3:
#cxo Automation
can provide foundation for customer self-service. Some customers may value
greatly, but others may consider a burden
A3:
#cxo If you
automate the human engagement element out of a process, you may be sabotaging
what customers value most. Careful!
A3:
#cxo Automation
can address quality elements: speed, consistency, and transparency. But don't
automate fundamentally bad solutions
A2:
#cxo If your
product is so-so, but service top-notch, I'll cut you some slack. If you show
steady prod-quality improvements, I'll stay
A2:
#cxo If I
purchase product from you, and it screws up once, & you take forever to
repair, your poor service has annihilated quality
A2:
#cxo Quality
intrinsic to entire customer value experience--products, services, etc. No
Solomonic cleaving them: all equally important
A1:
#cxo Quality is
each of you sweating the details of all the work that you personally have some
contribution to.
A1:
#cxo Quality is
NOT a class that your boss forces you to attend for 3 days and then never
mention again.
A1:
#cxo Quality
comes down to everybody vetting everything constantly at every level: inputs,
work in progress, deliverables.
A1:
#cxo Orgs that
emphasize quality take no excuses for shoddy work. Serious consequences for
inattention, negligence, & laziness
A1:
#cxo Quality is
obsessive attention to getting it right. Orgs focused on quality have culture
that incentivizes that obsession
1hr
to #CXO chat:
Build’g #quality
in the #custexp
w/ @JTweeddale http://bit.ly/1lWdiKB
#cxpa #mktg #custserv
Retweeted
by jameskobielus
RT
@IBMAnalytics @jameskobielus answers all your #bigdata Qs http://bit.ly/1jL4Efi
pic.twitter.com/ix0m2ppYH6 JK--Yikes! Not
literally ALL!
"Adopting
Information Governance in #BigData
Era" (http://ow.ly/vLyZQ ) JK--#IBM Redbook.
Download: http://ow.ly/vLzc8 #IBMDataMag
"Governance
Building Blocks" (http://ow.ly/vLxyp ) JK--Redbook: "#IBM Info Gov
Solutions." Download: http://ow.ly/vLxTX #IBMDataMag
"Open
Ecosystem Access" (http://ow.ly/vLx6G ) JK--Details on the new
IBM InfoSphere BigInsights Version 2.1.2. #IBMDataMag
"Data
Is Reality" (http://ow.ly/vLrHJ ) JK--Steve Adler says
"open data standards can help societies safeguard pasts+futures" #IBMDataMag
Catching
up on loose ends. I do this in bursts so that I'm never at wit's end.
RT
@kexpplaylist #kexp The Robots
by Kraftwerk from The Mix JK--1991. This remix gives the robot a warmer
personality and a bit of sly wit.
Machine
learning? Autom8ng log-data anlysis thru unsup/rnforce learn algos p1: http://ow.ly/vL7n0
p2: http://ow.ly/vL7oL
Mon #IBM qh
RT
@kexpplaylist #kexp Do I Have
the Time by Jessica Lea Mayfield from Make My Head Sing JK--2014. Wow. Electric
rush becomes this folkie.
RT
@kexpplaylist #kexp Make It Good
to Me by SJones & Dap-Kings from Dap-Dippin' With... JK--2002. By now, an
oldie...that sounds oldie-er
"Made
in #IBM Labs:
Dynamic Prioritiztion o Data in Cloud" (http://ow.ly/vL549 )
JK--QoS priorities can B modified per real-time/expected
I've
never once felt I needed to read the articles LinkedIn recommends. If perhaps
LinkedIn read my posts to LinkedIn, it might have a clue
Classical
radio playing "Lone Ranger" theme, "Hello Muddah Hello
Faddah," & somethng from Bugs Bunny. Glad they're finding apps 4 this
stuff
RT
@kexpplaylist #kexp Straight
Edge o th Blade by Jenny & Johnny from Im Having Fun Now JK--2010. Fine
exampl o Jenny Lewis sweet acid ooze
Hillary
handled th flying shoe incident like good stand-up comic, cracking wise and
absorbing the incident into the ongoing flow of her talk
"Artificial
Intelligence Is Next Frontier 4 Social Nets" (http://klou.tt/1rhgpdydveazl )
JK--AKA cognitiv computg, machine learning, image...
"It's
Official: Forrester Says Campaign Marketing Is Dead" (http://bit.ly/1sVna9e )
JK--They officially say it. It's not officially true.
RT
@manamica: "The reward of a
thing well done is to have done it." - Ralph Waldo Emerson
It's
funny how people seem to think "evangelist" is a nouveau term for IT
thought-leadership marketing. I've heard it for many years.
RT
@kaggle: New challenge w/350MM rows o
fascinating basket-level data. Predict which shoppers become repeat-buyers? https://www.kaggle.com/c/acquire-valued-shoppers-challenge …
I
have the utmost respect for people who deliver consistently without complaining
or making excuses. Just do the work, please.
Building
our pipeline of great contributors to #IBMDataMag.
Feels good to see all the great pieces queued up for posting.
Cool.
Weight's down to the lowest it's ever been in my adult years. Not doing
anything different. Just working hard & watching what I eat.
WSJ
Stanley Bing: "Social media is linking all human life on the planet into
one gigantic brain stem throbbing with unintelligible thought."
I
think Samantha Bee & Jason Jones should get Colbert's vacating slot on
Comedy Central. A dynamic married duo.
WSJ
says digitizing Vatican library a "mammoth task, involving 43 quadrillion
bytes." Just say 43 petabytes. I had to translate in my head.
Need
a new word to describe a new human-induced geological phenomenon:
"frackquake."
Anyone
who has watched Stephen Colbert for any time can see what kind of person he is.
Steps out of character often. Interviews, especially
OK,so
somebody several hundred years after Jesus' death says he mentioned a wife.
I'll accept that. Matthew, Mark, & Luke--what'd they know?
