Saturday, March 15, 2014

Aweekstweets March 7-15 2014: the week the wife and I crossed the big three oh

Curious why we're supposed to marvel at myth o St Patrick driving snakes out o Ireland. Burrowing serpents are good for soil. Eire needs 'em

 8m
One of most sobering aspects of looking through ancient biz cards from tech startups: most of those firms are now either kaput or acquired

 44m
Remember how folks were cool w/surveillance in Boston Marathon bombing aftermath? Likewise, now w/Malaysia Airlines incident. Less so re NSA

 1h
Did a pre-spring cleaning this morning down in my home office. I've collected 20+ years of biz cards. Dumped all but those with stories

 1h
Our 30th. Fun facts: 1) was also anniv of Liz & Dick's 1st nuptials & 2) also met my honey on set of Hollywood extravaganza we co-starred in

 9h
Several weeks running, WashPost's Saturday letters-to-editor page has been publishing anti-Beatles rants. My rebuttal? Stuff it, tin-ears!

What's especially stressful is trying to remember all these stress-reduction tips everybody keeps flooding the Internet with. UNCLE!!!!

 9h
RT @simonlporter: IBM's new #PaaS platform is #BlueMix. Join the beta (and survive the robot apocalypse): http://ibm.co/1gl0FRu  #devops

 25h
 Retweeted by jameskobielus

The Beauty of π (Pi) – An Iterative Calculation: http://bit.ly/OdlDv7  #abdsc via @analyticbridge
 Retweeted by jameskobielus

 22h
 Retweeted by jameskobielus

 21h
Smarticle! > #BigData Counting - How To Count A Billion Distinct Objects Using Only 1.5KB Of Memory: http://highscalability.com/blog/2012/4/5/big-data-counting-how-to-count-a-billion-distinct-objects-us.html … by @abramsm
 Retweeted by jameskobielus

 21h
A Survey of Dimensionality Reduction Techniques for #DataMining & #MachineLearning in High-Variety #BigData: http://arxiv.org/abs/1403.2877 
 Retweeted by jameskobielus

 24h
Malaysian authorities should enlist enterprise IT specialists in the search. If we can boil the Indian Ocean, plane might precipitate out.

National Pi Day. It's every American's patriotic duty to think & behave irrationally. Oh wait, that's what got us into this mess. Never mind

"17 areas to benefit from #bigdata analytics in next 10 years" (http://ow.ly/uAv6d ) JK--Great list of use cases & applications.

"#BigData & Privacy: An Uneasy Face-Off For Government To Face" (http://ow.ly/uAuNe ) JK--Read for the discussion of research projects

"Is Apache Spark the Next Big Thing in #BigData?" (http://ow.ly/uAtA9 ) JK--Gear to coming era of all-in-memory streaming architectures

Social media have nurtured a world culture that likes to be liked for liking to like stuff.

RT @kexpplaylist #kexp Monkey Gone to Heaven by Pixies from Doolittle JK--1989. "If the devil is 6 [sick 666] then God is 7 [is Heaven]."

RT @kexpplaylist #kexp Oh, Whisky! by Jimi Goodwin from Odludek JK--2014. Voice of the Doves. I love this one.


#BigData development? Don't understaff & overstretch your analytics development team (http://ow.ly/uAlNc ) Friday #IBM quick-hit

@BizIntelAnalyst No. It's a put-on.

This just in: Snowden says NSA knows where the Malaysia Airlines flight is and he has documented proof.

Missing Malaysia Airlines flight. No new facts since Day One. But that doesn't stop cable news from reporting pure speculation as news. 24x7

New #IBM jk blog: "Using #bigdata to tune the fusion of human physiology & machine physics" (http://www.ibmbigdatahub.com/blog/using-big-data-tune-fusion-human-physiology-and-machine-physics …)

"#IBM resrchr can decipher personality from looking @ 200 o yr tweets" (http://ow.ly/uz0jj ) JK--Honed on Twtr & IBM internal social net

Quantum uncertainty. I suspect it's God playing a shell game on us. We'll never see what's under the momentum & position shells at same time

RT @kexpplaylist #kexp Bread for Brat by The Hidden Cameras from Age JK--2014. This group puts out gorgeously orchestrated pop gems.

"Scaling #BigData Mining Infrastructure: Twitter Experience" (http://ow.ly/uyHA3 ) JK--Interesting one from SIGKDD Explorations journal

"5 things that'll remake #bigdata in next 5 years" (http://ow.ly/uyA5o ) JK--Hdp as platform, deep learning, quantified self, cloud, law

@ITAdvisorJD Apparently, you're under the impression that your contacts speak for IBM.

@batess Apparently, you didn't actually read my 140 characters.

"Real-Time Acoustic Processing Has #BigData Potential" (http://ow.ly/uyyM6 ) JK--ML analysis of ambient audio to predict near-futures

It can handle all data types, SQL & NoSQL, transactional & mixed analytics workloads:AND is fast & simple: #IBMBLU http://bit.ly/bluspeed 

U do dat data? Yes indeed, dude, I indubitably do! Do U?

Biz process optimization? Plugging a lean, mean analyzing machine into your manufacturing operations (http://ow.ly/uxSP9 ) Thurs #IBM qh

RT @kexpplaylist #kexp Heaven Knows I'm Miserable Now by The Smiths from Hatful of Hollow JK--1984. In Morrissey terms, "now" = "always."

Meet one of Network World's Fascinating 50: @LisaSeacat http://bit.ly/1hcReWf  #IBMMobile

The New Normal: Engaging with the mobile video-consuming consumer http://huff.to/1nkqzOH  #IBMMobile

I always thought "Westerns" were wrongly named. Never far enough west. Never saw Pacific Ocean. They were PMDs (Plains Mountains Deserts).

Rumor is that Christie will allow Tesla direct-sales in NJ only if owners forced to let filling-station attendants pump their electricity.

Often not better to challenge someone in the moment. Better to write a sharper point of view. A whole thought, uninterrupted, in response

Another crazy day. I begin and end each the same way. Clean desktop. I'm not one to let stuff fester in my workspace.

A8: #bigdatamgmt #BigData #Analytics integral cuz fraud doesn't announce itself. Must be discovered thru mining/monitoring transaction data

A7: At the end of the day by protecting the enterprise, this protects the integrity of the brand #bigdatamgmt
 Retweeted by jameskobielus

A7: Brand is important, but there are also economic impacts. 3.7T lost to fraud each year slows business expansion, jobs, etc. #bigdatamgmt
 Retweeted by jameskobielus

A7: #bigdatamgmt Counter #fraud measures, when incorporated into brand's customer experience, signify sound management. Brand confidence

A7: #bigdatamgmt Counter #fraud = trusted brand: weeding out fraudsters & freeloaders who raise costs that get passed onto legit customers

RT @mt_marko: A6: IBM and other vendors have built-in analytics capabilities in their SIEM products http://qub.me/  #BigDataMgmt
 Retweeted by jameskobielus

@IanGertler It's a risk analysis: risk of one terrorist slipping through vs. risk of millions of travelers getting super-annoyed.

A6: Visual representation can illustrate story behind data, and demonstrate linkages that are not obvious between entities #bigdatamgmt
 Retweeted by jameskobielus

A6: #bigdatamgmt Need visual exploratory & statistical tools in anti-fraud to sort through behavioral attribute data markers of fraud

A6: Entity, content, investigative, predictive and other analytics capabilities can addresses all phases of enterprise #fraud #bigdatamgmt

A6: #bigdatamgmt Predictive, graph, entity, & behavioral analytics are key to counter #fraud. Need anti-fraud SMEs/tools in data science CoE

A5: #bigdatamgmt Can reduce fraud false-positives if enlist expert human judgment vetting likely-frauds identified in data

A5: #bigdatamgmt Counter-fraud tactics must always weigh the risks of fraud against the risks of customer churn/anger from false-positives

A5: #bigdatamgmt Reviewing alerts quality and create accuracy metrics
 Retweeted by jameskobielus

A5: #bigdatamgmt Can never guarantee ZERO false alerts, Can reduce false-positives (but incr risk) by define more stringent fraud critieria

A4: Federal government is getting back 7.90 for every dollar of investigation spent for Healthcare claim #fraud #bigdatamgmt
 Retweeted by jameskobielus

RT @knowlengr: #bigdatamgmt Logging big part of detection / forensics. Amazon told r NYC Meetup their biggest DB:= billing / performance...
 Retweeted by jameskobielus

A4: .@InfinityAuto reduced referral of questionable claims 95% while processing legit claims to drive high client satisfaction #bigdatamgmt
 Retweeted by jameskobielus

A4: Orgs that take action to enhance their counter #fraud capabilities retain their customer’s trust #BigDataMgmt
 Retweeted by jameskobielus

A3: Give them a holistic view across the enterprise such that patterns and #fraud indicators are more easily identified #bigdatamgmt
 Retweeted by jameskobielus

A3: Incorporate the same counter #fraud tools across departments that enable more secure information sharing #bigdatamgmt
 Retweeted by jameskobielus

A3: Communication! Something that might not appear to be red flag for fraud for you might be for IT. #bigdatamgmt #security
 Retweeted by jameskobielus

A3: #bigdatamgmt Siloed depts can fight fraud by collab on operational predictive models & biz rules to spot & interdict fraud in time.

A3: #bigdatamgmt Siloed departments can fight fraud by pooling all customer profile, transaction, and interaction data.

A3: #bigdatamgmt Siloed departments can fight fraud by consolidating customer data marts into an enterprise data warehouse

A2: #bigdatamgmt In a #bigdata sense, "grass-roots vigilance" = event mining, interactive exploration, graph analysis.

A2: #bigdatamgmt Law enforcement teaches that fighting fraud is about unraveling a social network, node by node, thru sleuthing.

A2: #bigdatamgmt Law enforcemt can teach that top-down surveill has limited effectiveness against conspiracies. Need grass-roots vigilance

RT @BobGriffin_PS A1: Smarter system would recognize anomalies, patterns & entities that link a transaction to previous #fraud #BigDataMgmt
 Retweeted by jameskobielus

A1: Here's an interesting article about balancing that challenge http://bit.ly/KfzL5z  #BigDataMgmt
 Retweeted by jameskobielus

RT @mitchderman A1: Here's an interesting article about balancing that challenge http://bit.ly/KfzL5z  #BigDataMgmt
 Retweeted by jameskobielus

A1: #bigdatamgmt If, in background, you continuously mine transaction/interaction data for fraud markers, #custexp needn't suffer.

A1: #bigdatamgmt Build checkpoints into business processes that flag events indicative of fraud. Log/analyze those events in #bigdata

A1: #bigdatamgmt You need to leverage #bigdata & refine behavioral graph models that point fraud rings for special scrutiny.

"What is #NoSQL and how can it help my business?" (http://ow.ly/uvzzF ) JK--This overview is a keeper for tech professionals.

"Tips 4 Better #BigData Presos" (http://ow.ly/uvyTo ) JK--Best "why presos fail": "overconfidence in one’s ability as an awesome analyst"

@loserboy It's nice to hear Mark Kozelek "sell out" at least once in his career. I respect sell-outs if they're listenable + insanely brazen

"Top Ten Reason To 'Kaggle'" (http://ow.ly/uvxD4 ) JK--Good list, but he actually lists 12 reasons. And title has single-plural mismatch

"Some Useful Machine Learning Libraries" (http://ow.ly/uvxdz ) JK--Useful if you're a hardcore data scientist.

@loserboy Whew! Relieved. Depressive grooves become him. But this is a cool track. I like the ballsiness of putting in a Kenny G-like sax

RT @kexpplaylist #kexp Ben's My Friend by Sun Kil Moon from Benji JK--2014. This is Sun Kil Moon? Sounds sort of happy. Wazzup?

"Interactive Graphic: How Big Is #BigData?" (http://ow.ly/uvv34 ) JK--Hmm. Per domain, plots data dimensionality against sample size

"Life Cycle of Data Science Projects" (http://ow.ly/uvuFu ) JK--Good outline by Vincent Granville.

This morning's blogosphere trawl on #bigdata is slim on substance. Too many facile renamings & bogus "we discovered it first" pronouncements

RT @kexpplaylist #kexp Oh Lonesome Me by M. Ward from Hold Time JK--2009. Peppy Don Gibson classic re-imagined as heroin addict suicide note

Post by @scharffc How Mobile Computing is Enabling Personalized Learning http://ibm.co/N5ZA8H  #IBMmobile

Experience optimization? #BigData framing th engagement with art & culture p1: http://ow.ly/uvey8  p2: http://ow.ly/uvezz  Wed #IBM qh

Good day of chatting with analysts. Many of them I know from long back. Hope they can contribute to #IBMDataMag

"Machine learning in 10 pictures" (http://www.denizyuret.com/2014/02/machine-learning-in-5-pictures.html?m=1 …) JK--From the POV of a data scientist.

RT @Datamation: .@jameskobielus shares his expertise about #bigdata during our live hangout. http://bit.ly/BigDataHOA  pic.twitter.com/ElKNI0FQ4k

How do you fight fraud in a #BigData World? Join #IBM and #IDC for the next #BigDataMgmt chat on 3/12 at 12pm ET http://ibm.co/1gW89vV 

Small world. One o my new contributors to #IBMDataMag once dated one o my favorite #KEXP DJs. Won't say who dated who. But they're both cool

"Getty makes 35M images embeddable to gather data, monetize usage" (http://ow.ly/utuWY ) JK--Free. Incl photog attrib + link 2 Getty site

Before long, we'll stop calling it "machine data" and start referring to "thing data." Feels inevitable.

Drafted latest #IBM blog: "Using #BigData to tune the fusion of human physiology and machine physics"

Geospatial analytics? Statistical lenses strip darkness from the distant cosmos (http://ow.ly/usHzm ) Tuesday #IBM quick-hit

RT @kexpplaylist #kexp There She Goes by The La's from The La's JK--1988. Love this original. Also love Sixpence None the Richer's cover.

If you look at photos of American men from the early 1970s from the head up, they resemble American men from the 1870s. Shaggy mops & beards

Rewatching "The Paper Chase" (1973). Still the most realistic college movie ever. 5 years later: "Natl Lampoon's Animal House." The funniest

Don't ever abbreviate "root cause analysis" to "root canal-ysis." I'm just sayin'.

Cumberbatch: especially cumbersome batch-processing job. Benedict: benediction U say before hitting the "process" button. Please don't crash

Hmmm. #kexp will be at the center of my quick-hit for next Monday. A "recommendation engines" theme. Why'd it take me this long?

RT @wfuessler: A8. High performing #CFO known as the 'Performance Accelerator' are 2x likely to identify revenue growth opportunities. #CXO
 Retweeted by jameskobielus

A8: #cxo Analytics take #custexp out of the "qualitative" initiatives into the quantitative. Satisfaction as a measurable asset.

A8: #cxo Analytics help #CFO to track hard-dollar KPIs on #custexp needed to justify programs to CEO, Bd of Directors, & investors

A7: #cxo Less expensive in long run to retain customers than win back churned customers. Improved #custexp reduces churn.

A7: #cxo Implementing automated #CEM next best action in multichannel equals handle more transactions/interactions w/lower overhead

A7: #cxo The "money" nexus: better experience = more retained/acquired customers = larger share of "pocket" = larger sales/rev

RT @clearaction: A6: ROI of #CX is often in savings: lower costs, shorter time, less waste due to better alignment to customers #cxo

A6: may sound lowtech - just use #BigData and demonstrate correlation between repeat customers #cx investments and profitability #cxo
 Retweeted by jameskobielus

A6: #cxo #CFO buy-in to #custexp investments come from hard-dollar paybacks in churn reduction, sales/rev boost, & cost control

A6: #cxo Get #CFO buy-in into #custexp investmts by stress "money" over "love" ROI metrics, per this dude: http://bit.ly/1kb2g0N 

A5: #cxo Secure #Mobile #digital channels enable anywhere/time/yone fin report+tracking. Accelerates closed-loop biz-proc optimization

A5: #cxo Digital channels enable #CFO org to automate real-time rollup/tracking of financial metrics from operations.

A5: #cxo When used B2E to streamline #CFO fin reporting/analysis, digital channs squeeze manual costs/delay/errors from process

Q4: By funding digital strategies that maps their customer journey #cxo
 Retweeted by jameskobielus

A4 If the CFO is providing org insights to support decisions by CMO COO CEO then better than most enterprises. #cxo
 Retweeted by jameskobielus

A4: #cxo #CFO helps prioritize the degree to which biz manages data 4 monetization purposes vs. #CXO multichann mktg etc.

A4: #cxo #CFO role in transform #custexp always indirect--eg, clarifies biz-perf metrics 2 B optimized in front- v. back-office A/B tests

A3: CFO must have laser focus on value creation and monetization #cxo
 Retweeted by jameskobielus

A3 The power of the CFO to illustrate a story backed by financial data is business advice. #CXO
 Retweeted by jameskobielus

A3: #cxo If nothing else, the #CFO needs to provide the core/canonical fin perf analytics mart all LOB functions rely on

A3: #cxo #CFO needs to apply the ROI "nasty stick" to help C-level prioritize disparate customer-centric data/analytics initiatives

A3: #cxo To elevate #CFO into provider of #biz advice, CFO needs to assist CEO in measure bottom-line contrib of cust engagemt projects

A2: RT @hammnick: Does yr biz have “make it right” fund that empowers front line employs & managrs w/budget to ensure over the top #cx? #cxo

A2: #cxo Most important role of #CFO in closing "customer expectation-execution" gap is ensuring biz has funds to drive execution

A2: #cxo #CFO needn't be a jack of all trades. #CMO #CIO #COO & other peer execs all collab in closing expectation-execution gap

A1: #cxo Customer as core stakeholder means that churn & abandonment are most critical showstopper business issues.

A1. The Enterprise must show customers & employees that they care and trying to think more like a startup! Agility & Transparency! #CXO
 Retweeted by jameskobielus

A1: #cxo Companies will begin to make it easier 4 customers to share prod/service feedback and they will ACT on it!
 Retweeted by jameskobielus

A1: #cxo Customers have always been the core stakeholders. The value U render 2 them is paramount reason 4 being in business.

A1: #cxo Future of enterprise will be much like past: if U don't keep customers uppermost, you'll lose them. And you'll fail

A1: #cxo Customer is stakeholder. That means you're accountable to them--ie, U need to explain/justify/report to them regularly

@IBMbigdata Happy to be pitching in my thoughts on today's #CXO tweetchat

Reading GCarlin's posthumous autobio, you can't help noticing how intensely careerist he was. Even his "hippie" metamorphosis was calculated

"What Is #Hadoop Exactly? A Cynic's Theory" (http://ow.ly/uqzGF ) JK--Eric Kavanagh asks what I've been asking since my Forrester days

New #IBM jk #AnalyzingMedia column: "Detecting the Signs of Drama and Drift in Customer Loyalty" (http://ow.ly/uqy6a )

"Machine Learning a Major Part of #BigData" (http://ow.ly/uqx4u ) JK--"Crown jewel" in big data project. IBM

"Analytics & Belief: Th Struggle 4 Truth" (http://ow.ly/uqwyQ ) JK--Cool epistemological dissection. Appeals to my inner philosophy geek

"Predictive model used in air traffic cancellator" (http://ow.ly/uqvz5 ) JK--Nice piece on OR (mathematical optimization) in data science

"Sometimes outliers are real data" (http://ow.ly/uquI1 ) JK--When is outlier a fluke vs. valid evidence of a deeper pattern to explore?

"Why Apache Spark ...." (http://ow.ly/uqub5 ) JK--Ignore the hopeful eager-beaver headline. Good informational article, nevertheless.

"Big Data vs. the Data Warehouse" (http://ow.ly/uqtpH ) JK--Steve Miller dissects Bill Inmon's retro-defensive definitional turf-guarding

CountDwn 2 #CXO: The #CFO: Jack of all trades http://ibm.co/MUadeq  w/ @wfuessler et moi CU in 1 hr 20 mins! #custexp #finserv #cxpa

Pleasant morning of invitation sending. I read as much as I write. Reaching out to people that I read & value for thought leadership.

Healthcare analytics? Sensing, mapping, & mining the mystery of the brain p1: http://ow.ly/uq7tT  p2: http://ow.ly/uq7vJ  Mon #IBM q-h

Just saw somebody describe self on Twitter as author of a "fiction novel." One of those retronyms, like "postal mail" or "acoustic guitar"

It was, for me, a relatively social-media-free weekend. By choice. Occasionally, I get tired of seeing my own words.

What is this "the new black" stuff all about, anyway? Some ingredient that goes well with everything? How about a sense of perspective?

RT @kexpplaylist #kexp On the Beach by Neil Young from On the Beach JK--1974. One of his cool slow-burn songs. Quiet, intense, brooding jam

Pointedly NOT following moves of this latest yoga instructor. Seem calculated to cause injury. The pits.

Made up for the lost spring-ahead hour last night by dreaming I had slept an extra hour. Virtually caught up on my Zs.

"Doonesbury Flashback" now running the primitively drawn early-70s strips. Perhaps Trudeau needs to retire the strip for good. Played out.

"The Francis Effect"? So what!? Going to take a major shift in church policies to close the cultural rift alienating the mainstream laity.

DC NBC affiliate has "Meet the Press" moderator David Gregory at the news desk. That dude's tall. I've seen him at the airport.

Pleasant enough Sunday morning. The usual light breakfast followed by moderate workout. Did all the chores yesterday. Later? Nothing.


Good night of reading, drinking wine, eating popcorn, lying on back. I don't watch much TV anymore. Barely miss it.

Nice to get out there and wash the cars. Removed three months of cold grimy slop. Gleefully watched the snowpiles dissolve.

@odbmsorg Of course it's free. I never pay a dime.

Everybody's out in force, enjoying the mild weather. Never mind the ubiquitous puddles & mud from snowmelt. A lamb-ish Saturday of shopping

Looking for fresh #bigdata thought leaders to become regular contributors to #IBMDataMag. Inquire within (i.e message me for details)


Man, I been wasting daylight all winter long. Glad we'll finally get around to saving some for a while. Need 2 bank it again for next winter