Friday, January 24, 2014

Aweekstweets January 18-24 2014: the week it stayed subfreezing

"#Hadoop 2.0's deep impact on #bigdata techs" ( ) JK--Nicole Laskowski rounds up the analysts

"#BigData Debate: Will HBase Dominate NoSQL?" ( ) JK--Will? It's not dominating it now. Debate premise is a bit off.

Catch me #IBM @fhalper #TDWI Krish Krishan, et al on #BigData webinar on Tues Feb 11, 12-1pm EST ( ).

Video: The Retail Equation predicts and shapes shopper behavior with #IBM #PureData for Analytics ( )

Stupidest LinkedIn message is "a LinkedIn member viewed your profile." Exactly what am I going to do with that information?

U can tell IoT is hot by the growing surge of cautionary pushback, eg 

#CXO is back Mon 12 ET! #Data 4 #CX Advantage w/ @StephanieThum, @WilliamMcKnight y yo Señor Diego Kobi  #bigdata

Occasionally a typo coins a gr8 new word. Discussing historical archives, I was thinking of "yesterday's data" & it came out as "yesterdata"

"How Machine Learning Could Result In Great Applications for Your Business" ( ) JK--Good discussion of specific apps.

Meaty metadata? Data variety leads to metadata viscosity p1:  Friday #IBM quick-hit

RT @kexpplaylist #kexp Tomorrow is a Long Time by Phosphorescent from Sweetheart 2014 JK--A fine cover of a simple soulful Dylan classic.

"Infinity AR: We'll fulfill sci-fi promise of augmented reality" ( ) JK--Curious why they demo privacy-invading AR app

Ejected DVD of Leonardo Di Caprio "The Great Gatsby" a half-hour in. Hyperstylized junk. Hiphop in soundtrack of the Roaring '20s? WTF?!

RT @caro: I'd be OK if law pardoned Rob Ford & Justin Bieber on the condition that they recorded duet of "Paradise by the Dashboard Light."

Have drafted 4 of next week's 5 quick-hits. Now I'll breathe and knock off for the evening.

"Big Data & Clinicians: Review on State of Science" ( ) JK--In-depth discussion by Stanford School of Medicine rsrchers

"Framework to build logistic regress model in rare event population" ( ) JK--Interesting step-by-step from data scntist

Responded to email interview on #bigdata by ODBMS Industry Watch.

RT @kexpplaylist #kexp You Are the Generation That Bought More Shoes & You Get What You Deserve by Johnny Boy JK--2006. Gr8 song. Long title

Drafted latest #IBM AnalyzingMedia blog: "Pay Your Customers What They're Worth to You"

"How Government Can Make Open Data Work" ( ) JK--Good detailed practical ideas.

"The internet of things needs a new security model. Which one will win?" ( ) JK--It doesn't have an old model. Up4grabs

New #IBM jk blog: "#BigData overkill can stunt scientific rigor" ( )

"What’s the Lift of Your Churn Model?" ( ) JK--Popular pickup line in bars frequented by data scientists.

"Three myths about data scientists and #bigdata" ( ) JK--New word (new to me at least): "polyvalent." Means "versatile"

RT @kexpplaylist #kexp Dreams by The Electric Peanut Butter Company from Trans Atlantic Psych Classics Vol. 2 JK--2013. Fleetwood Mac cover

"Lenovo Plans 2 Acq #IBM x86 Svr Biz" ( ) JK--IBM keeps Sys z, Powr Sys, Stor Sys, Pwr-bsd Flex svrs, PureApp, PureData

Context accumulation? Narratives drive home relevance of stat models p1:  p2:  Thurs #IBM qh

In the Middle Ages the English referred to the Black Death as the "Great Death." I find that even more chilling.

RT @kexpplaylist #kexp Natural One by Shearwater from Fellow Travelers JK--2013. Great cover of Folk Implosion's 1995 original.

"Apache Spark: The Next #BigData Thing?" ( ) JK--I'm most interested in its streaming model. See mid-article.

"The future of storage: disk-based or just discombobulated?" ( ) JK--Discusses recent M&A in the storage market.

"New Techniques Detect Anomalies in #BigData" ( ) JK--Discusses anomaly detection relevant 2 IoT/machine data analytics

"Can #BigData & SQL Get Along?" ( ) JK--Discusses pros and cons of SQL in a NoSQL environment.

Lookee here: 15 balmy degrees in the middle of the afternoon. Well, at least it's sunny. Cold consolation.

New jk #IBM #Dataversity article: "The Beauty Metric: Choosing the Best-Fit Advanced Analytic Algorithms" ( )

A8: u don't need a schema 2 enforce security, u need a better security tool that monitors & secures Hadoop/NoSQL #bigdatamgmt
 Retweeted by jameskobielus

A8: #bigdatamgmt You can have flexibility 2 use any #bigdata platform U wish. As long as plats individually & jointly are secure

#BigDataMgmt A8: #Hadoop is not the issue. The lack of a holistic approach to #BigData #Security & #Privacy is. End the Patchwork!
 Retweeted by jameskobielus

A8 Security tech needs 2 adapt 2Hadoop/NoSQL - not the other way round. #bigdatamgmt
 Retweeted by jameskobielus

A8: @merv suggests AAAA approach: authentication, authorization, audit, anonymization  #bigdatamgmt
 Retweeted by jameskobielus

A8: #bigdatamgmt Sure. But that demands that #privacy #security be addressed ALL the time into ALL planning and ops for ALL #bigdata

+1 RT @DCorrigan: A7: Also need to identify sensitive data, and know it's lineage and what it's used for #bigdatamgmt
 Retweeted by jameskobielus

#bigdatamgmt A7 Centralization => Risk of bigger failures, bigger bullseye on back
 Retweeted by jameskobielus

A7: #bigdatamgmt If #Hadoop deployed for ETL into DW augmentation role, & if DW holds PII, Hdp ops must also protect #privacy

A7: #bigdatamgmt If data in #Hadoop clusters not PII, I have no concerns. If it's PII, is there implied consent (eg social media data)?

#BigDataMgmt: A7: #Hadoop was not Designed for #Privacy. Vendors commercializing this IP must address it. #BigData will #fail without it.
 Retweeted by jameskobielus

A7: Partly No, Appropriate configurations & Apps done right could provide enterprise grade security in #hadoop #bigdatamgmt
 Retweeted by jameskobielus

A7: #bigdatamgmt My concern is not with security of #Hadoop tech. It's with business' lack of data stewardship controls around Hdp data

A6: #bigdatamgmt Sheer proliferation of DBs containing PII is one of the biggest barriers to #privacy compliance. Find it all first.

A5: #bigdatamgmt The bigger the business is, the more likely to be target of attack. But zombies target businesses at all scales.

A6: #bigdatamgmt Lack of unified PII governance under #MDM is big barriers to #privacy compliance. PII data not yet harmonized

A6: #bigdatamgmt Failure of a company to have chief privacy officer or privacy data stewards is a tech barrier. Who's responsible

A6 - fragmented security technologies 4 each data platform. U need 1 security technology 2 monitor data in Hadoop, RDBMS, NoSQL #bigdatamgmt
 Retweeted by jameskobielus

A6: Different IT infrastructure & associated data policy makes it harder to achieve privacy compliance #bigdatamgmt
 Retweeted by jameskobielus

A5 Orgs need to work on assumption they've already been breached. Korea = inside job; Target unnoticed for too long, etc. #bigdatamgmt
 Retweeted by jameskobielus

A5: Biggest data security threats come from inside, report says  #bigdatamgmt
 Retweeted by jameskobielus

A4: Remember that our brain wiring fights us here. I wrote chapter on this here  #bigdatamgmt #bigdata, #privacy
 Retweeted by jameskobielus

A5: Accidental disclosure tops hackers! #Forrester says 36% of breaches are internal  #bigdatamgmt
 Retweeted by jameskobielus

A5: #bigdatamgmt Accidental disclosure is always a threat. The more complex & dynamic a biz, more likely someone is to slip up

A5: #bigdatamgmt Cluelessness. It's too easy for biz to take for granted data they hold & misuse it without awareness of consequences

A4 - The punishment will be a silent killer - how many customers avoided you because of privacy issues? #bigdatamgmt
 Retweeted by jameskobielus

A4: Using the power you hold as a consumer is the best way. Example: Target, you won't find me shopping there any time soon. #bigdatamgmt
 Retweeted by jameskobielus

A4: #bigdatamgmt Shaming is a punishment in a reputation-sensitive economy. May not be easy to quantify the bottomline impact, though

A4: Target had to make an official statement, as quarter earnings materially affected. #bigdatamgmt
 Retweeted by jameskobielus

@IBMbigdata: A4: 2013 Ponemon Institute Data Breach Study says avg cost of a breach to a US org is $5.4M. #bigdatamgmt
 Retweeted by jameskobielus

RT @IBMbigdata A4: Here’s link to Ponemon Cost of Data Breach Study  #bigdatamgmt
 Retweeted by jameskobielus

A4: They lose #trust, loyalty, and instill fear in the consumer. #bigdatamgmt Still afraid to shop at #Target...
 Retweeted by jameskobielus

A4: #bigdatamgmt If by "punish" you're referring to expense of privacy litigation, damage control, & retooling, many biz have suffered

A4: #bigdatamgmt Good question. Has any privacy-deaf business ever suffered significant customer churn? I haven't heard of any.

A3: #bigdatamgmt #Privacy is #CX concern just as much as a #legal concern. You don't want to creep out customers w/ privacy invasions

A3: Orgs should rely on policy & law as a MINIMUM! Great orgs go beyond. #bigdatamgmt
 Retweeted by jameskobielus

A3 No! Work with customers to set reasonable policy, limits, uses. And be transparent. #bigdatamgmt
 Retweeted by jameskobielus

A3: #bigdatamgmt If you stay in customer comfort zone #privacy, you're mitigating risk of lawsuits. Be more stringent than law requires

A3: #bigdatamgmt Your org should consider customer "experience" sensitivities, PLUS law/regulations, in setting #privacy practices.

A2: #bigdatamgmt IMHO, users in online society should own/withhold some PII. It's a monetizable personal resource.

Individuals own their Digital Identities, but must be ever-vigilant in managing and protecting them to maintain any #Privacy #bigdatamgmt
 Retweeted by jameskobielus

RT @DCorrigan: a2 things are changing in EU - consumers own data, and have right to have it deleted.  #bigdatamgmt
 Retweeted by jameskobielus

A2: #bigdatamgmt If consider who SHOULD own PII, some feel individual should. But when gen'd in transactions, shouldnt biz share ownership?

A2: #bigdatamgmt The matter of data ownership is also obviously jurisdiction-specific. But laws may be vague in some places. Risks abound

A1: #bigdatamgmt The notion of "implied consent" must be considered on some types of data. HIre a good lawyer to advise you.

A1: #bigdatamgmt If the data you're collecting is NOT PII, there may still be relevant restrictions (eg regarding transborder data flows).

A1: #bigdatamgmt If the data is personally identifiable info, you need to consult relevant laws before collecting, managing, or using

Agreed! RT @IBMAnalytics: Grt point RT @Natasha_D_G A1: Social data = public data and thus orgs can collect at will #bigdatamgmt #privacy
 Retweeted by jameskobielus

A1: Open public data and data with waived rights (certain #CreativeCommon ) should float without consent #NotRelevantToNSA #bigdatamgmt
 Retweeted by jameskobielus

A1: #bigdatamgmt Depending on laws of the jurisdiction you operate in. The safe answer is "never." Get consent 1st before you collect PII

@IBMbigdata As always, I stand ready to tweet with all the pretentious but concise erudition that I can muster.

Maximize the value of your Systems of Engagement w/ IBM's Engagement #Analytics,  #bigdata #socbiz #ibmconnect

Apple's stupid "Lemmings" commercial from the 1985 Super Bowl ( ). Did Disney really license "Heigh Ho" for this?

Whoa! I just found a legitimate reason to reference Schopenhauer in a blog. All my smartypants pretentious erudition rising to the surface.

Ah yes, coordinating #IBM-wide response 2 an analyst firm. Not dissimilar 2 coordin8ng analyst-firm-wide respns 2 #IBM, which I've also done

RT @kexpplaylist #kexp A New England by Billy Bragg from the Peel Sessions JK--1991. This is his rawest and best version of this great song

NoSQL? The architecture that's still curiously absent p1:  p2:  Wednesday #IBM quick-hit

Hard to take an interest in the upcoming Winter Olympics. An armed athletic encampment under the control of homophobes? Go to hell, Putin!

Aging in your career. You know it's happening when people (and you yourself) routinely refer to you as a "veteran." Yikes!

Ever look at a newly snowy landscape and immediately lose memory of how it appeared pre-snow? You wonder if it's the same place underneath.

Crazy-busy day today. But made measurable progress on everything. That's all I seek. Now: freshfallen snow to remove.

"3 innovs 2 xform mob data analytx" ( ) JK--"Once bridg digital + phys, no way restrict possibs o val data & event gen"

"The number crunch: Will Big Data transform your life?" ( ) JK--Closes w/discussion of #IBM "computational creativity"

Drafted next #IBMDataMag article: "Truth, Beauty, Authority, and Efficiency: When to Place Faith in Statistical Models (and When Not)"

"Will these tiny computers herald the arrival of Internet of Things?" ( ) JK--Teensy-weensy platforms for thing devel.

"DIY Internet of things: Ultimate maker project" ( ) JK--"Homebrew" thing development? Who'll be the Jobs/Woz of IoT?

"2014: An Internet of Things Odyssey" ( ) JK--Fun fact: the monolith was God's lost smartphone calling home.

Ze snow she is falling!

RT @kexpplaylist #kexp Transmission by Joy Division frm Substance JK--"Radio live transmission"? No. Studio recrdng. Duh. Dance dance dance!

WashPost article about SilVal firms decamping 4 DC area. Focuses on wireless vendor. Yeah we have that depth. I was w/ WL tool vendor in 90s

Trying to think of a cute hype-y nickname for the coming snowstorm. Hmm. BlizzKrieg. Yeah. That's the ticket!

RT @kexpplaylist #kexp Blue Moon by Beck from Morning Phase JK--2014. Wow! Fresh Beck! And it's good stuff (naturally).

Waiting for the snow to begin falling. It won't be proverbial "snow day" for me. I still work--and doubly so--i.e., my job + the shoveling

RT @kexpplaylist #kexp Sleeping Gas by The Teardrop Explodes from Kilimanjaro JK--1980. Liverpool/Julian Cope. Similar 2 Echo & The Bunnymen

RT @kexpplaylist #kexp Silver Timothy by Damien Jurado from Brothers & Sisters of Eternal Son JK--2014. Odd, endearing voice like Neil Young

Security of #bigdata? Shoddy lifecycle management is ironic data security p1:  p2:  Tues #IBM qh

Mahinder told the party last night that he was impressed with my prior knowledge of Turkey's ancient history. Yes, I'm a total history geek.

"Internet of Things: Do Customers Have Say in Who Owns Their Data?" ( ) JK--Misleadingly implies B2C is only IoT domain

Right now, Weather Channel says 57 ° in NoVa. Tomorrow they predict 5-7 inches of snow. I'm sure the 57-5-7 stat correlation is significant

When some1 sends me list of tech Qs outa nowhere, saying "I look forward to yr response," they can look forward to eternity. Wot in it 4 me?

"Making sense o social nets, social graphs, collab, & cooperation in new way o work" ( ) JK--"Leanership"? Oh, brother!

Saw DVD of 1999 film "Angela's Ashes" about desperately poor Irish people pining for some magical fantastical wonderland called "America."

The Box Tops "Cry Like a Baby" ( ) JK--Recorded in Memphis. From Memphis. #2 on charts on 4-4-68.

Louis Armstrong "Ain't Misbehavin'" ( ) JK--One of biggest hits the year MLK Jr was born. 1929.

Otis Redding "Sitting on th Dock of Th Bay" ( ) JK--#1 song on 4-4-68. Died 10-10-67. Plane crash Lk Mendota Madison WI

10,000 Maniacs "A Campfire Song" ( ) JK--1987. Cool duet of Natalie Merchant & Michael Stipe

REM "Begin the Begin" ( ) JK--"Let's begin again, like Martin Luther Zen." Gr8 nonsense. I met these guys in Madison.

Drafted my latest #IBM #Dataversity blog: "The Beauty Metric: Choosing the Best-Fit Advanced Analytic Algorithms"

Drafted latest #IBM blog: "#BigData Overkill Can Stunt Scientific Rigor"

RT @kexpplaylist #kexp After the Disco by Broken Bells from After th Disco JK--2014. JMercer boasts 2 awesome & distinct groups (this+Shins)

"What analytic managers can learn from neural nets" ( ) JK--"Neural mgmt....all data scientists perform all functions"

"Tim O'Reilly on open data" ( ) JK--"When cost is low enough, it ...create[s] many of the same conditions as a commons”

"Human Analysts at Superhuman Scales" ( ) JK--Discusses their graph analytics programming framework for genomics

"PredMod w/#BigData: Bigger Really Better?" ( ) JK--“Crtn telling behavs may not B obsrvd in suff #s w/out massiv data”

"Markov Logic Networks" ( ) JK--"Situations where outcomes partly random and partly under control of a decision maker"

"How NLP Makes Lives Easier ( ) JK--"Ultimate goal o NLP is 'do away with computer programming languages altogether'"

Advanced visualization? Using visuals to discriminate among machine-learning algorithms ( ) Monday #IBM quick-hit

RT @kexpplaylist #kexp Red Eyes by The War on Drugs from Lost in the Dream JK--2014. These guys are good. Real solid rock with depth.

Weirdly, my 1st (crappy) job out of grad school gave me a useful skill. Reader in Wisconsin Newspaper Assn clipping bureau. Speed-gleaning.

Dinner party last night w/friends Mahinder & Harban who we met on Turkey trip last month. Reminisced about recent memories of ancient ruins

Saw article about how data, supposedly, is "evil." Article discussed how practices (collecting+using data) might qualify. Data itself? No

Congrats to Seattle and Denver. Super Bowl parties in both cities will be historic. Legal doobies will be lit to celebrate.

OK you numchuk ex-jock broadcasters: testosterone-fueled shouting of football cliches & truisms is NOT the same as analysis.

Sandy aid. Was it Christie's slush fund to reward NJ mayors who supported his re-election bid & punish those who didn't?

"Punk"? "Post-punk"? For the sake of exhaustive categorization, I'm going to insist that we label all rock before 1976 as "pre-punk."

"CBS Sunday Morning" on autonomous vehicles. These will offer unparalleled mobility for the disabled. Wheelchairs will dock inside.

"CBS Sunday Morning" examination of doodling. It's just a tactic to stave off boredom inflicted by people telling you nothing new or useful.

Reading Nicholas Ostler's excellent "Empire of the Word: A Language History of the World." Especiialy enjoy his theory of language spread

I love the elegance of Windows' design, best exemplified by the command "Force Shutdown."