Saturday, March 22, 2014

Aweekstweets March 15-22 2014: the week I typed these very words you’re reading now…wow!

RT @kexpplaylist #kexp I Can't Stand the Rain by Brian & Tony Gold JK--1997. Excellent reggae-funk hybrid.

Probably going to have dinner tonight at a place that looks like a hybrid of Thai restaurant and sports bar. See how that works out.

Reading collection o Nabokov short stories arranged chronologically starting in his early 20s. Exceptionally concise, contoured observations

"Working class." I've never understood the term. The only people it doesn't include are those who aren't working and don't care to.

Didn't realize this bit of music trivia: Andy Summers of The Police was briefly a member of The Animals in 1968.

So "Prickly City" is conservative. Big whoop. "Doonesbury" is liberal. I don't care. Just make me laugh. If comics don't do that, dump 'em.

US mints its first dome-shaped coin? Is vending-machine compatibility no longer a consideration at the Bureau of Engraving & Printing?

I tweeted my ass off while vacationing in Turkey a few months ago. Erdogan is an idiot if he thinks banning it is consistent with democracy.

Geez, can't believe I'm caught up on developing all my presos thru June. Fortunately, several common themes. I reuse slideware if possible

Watson Mobile Challenge to Usher in New Era of Development  #IBMMobile @momoboston

Everybody who follows me on social media & wants to know if IBM is hiring. Yes, we're always hiring. Please use the usual channels for that

Frenetic multitasking today. Finished the quick-hits & a presentation. Almost forgot lunch. Almost.

New #IBM jk blog: "Plugging a lean, mean #bigdata -analyzing machine into manufacturing" ( )

Apache Tez ( ) JK--Evolves YARN. Generalizes MR to execute a complex DAG (directed acyclic graph) of tasks. Now in v0.3

RT @kexpplaylist #kexp Sympathy For The Devil by Rolling Stones from Beggars Banquet JK--Somebody please explain "troubadours...Bombay" ref

"Using #bigdata 2 ID triple-neg breast, oropharyngeal, & lung cancers" ( ) JK--ML & pattern recog on MR images/biopsies

"#Hadoop Alternative Hydra Re-Spawns as Open Source" ( ) JK--Batch and streaming? Sounds more like a Spark alternative.

"Data’s Treacherous Journey from Screen to Mind" ( ) JK--That app generating social net from email history is way cool

"3 Data-Driven Nuggets From Nate Silver’s New News Site" ( ) JK--Stat anlysis o pres elect, fitness actvty, health news

"9 Types o Data & How Should B Used" ( ) JK--Not sure I agree with trustworthiness ranking, but worth pondering.

"Crime fighting with #bigdata weapons" ( ) JK--Shiver to think how predictive policing can be abused 4 social profiling

"#IBM Seeks Patent on Software that Incorporates Human Emotion" ( ) JK--Digitizing physio inputs 2 determine emotion

Big #BI? #NoSQL's role in decision-support applications ( ) Friday #IBM quick-hit

This video is the best news summary of the vanishing Malaysian airline story ( …)

"Pred Analytics Marketplaces: Viable Solution to Data Scientist Shortage?" ( ) JK--Discusses that + 2 other approaches

"Use Data to Tell the Future: Understanding Machine Learning" ( ) JK--Wired does good overview for general readership.

"8 Ways #BigData & Analytics Will Change Sports" ( ) JK--"Moneyball" is the least of it.

Storage-happy PB-cap suprcomputrs due in '15 ( ) JK--Memory/flash arrays. Wrangler (TACC/UT-Austin) & Comet (SDCC-UCSD)

"Trucking benefits from IoT data but struggles w/worker resistance 2 analytics" ( ) JK--Aversion 2 intrusive monitoring

"#NoSQL and Log Data Jump to Forefront" ( ) JK--Discusses its role in IoT sensor data. Focuses on aviation industry.

"#IBM Launches New Sw & Consulting Svcs to Help Orgs Tackle $3.5 Trillion Lost Annually to Fraud & Financial Crime" ( )

RIP Patrick McGovern. IDG founder & chairman was pivotal figure in development of global IT trade press, analyst, & market research sectors

Very nerve-consuming morning o schedule-rejuggling to fit in a speaking engagement. Collaborate 2014. Apr 7-9. Las Vegas. #BigData Ecosystem

Drafted next #IBMDataMag column: The Hidden Biases that May Cloud Cognitive Computing

Open data? The standards imperative ( ) Thursday #IBM quick-hit #IBMDataMag

Fraudsters can’t hide behind #BigData and a holistic approach to combat the $3.5 trillion challenge  #counterfraud

Imagine Collaborating with #IBM Watson? You Can  #IBMMobile @RitaJKing

"Data Wonks Swap Tips" ( ) JK--"Universal truth"? Unh-uh. Erroneous POV on present. #BigData not new. Knowledge abounds

We wonks are a distinct subculture among the geeks. We communicate via wonkie-talkie. We hang out down at wonkie tonks & wonk the night away

What's the max % of foreign-derived words in a sentence before it is no longer considered "plain English"? Is "Beowulf" plain enough 4 ya?

I wonder if the "chicken from hell" tasted like chicken? How did being raised amid fire & brimstone affect its flavor?

Dialogue from WashPost politician profile: "I'm Al Franken, junior senator from the state of Minnesota, clothed in immense power."

Rewatched "The Heart Is A Lonely Hunter" (1968). Made me look up whatever became of Sondra Locke. Alan Arkin & Stacy Keach, I know.

A8: #bigdatamgmt Also, recognize that some new unstruct data (e.g., social sentiment) may not require industrial-grade governance.

A8: #bigdatamgmt I'd also point to need to keep an eye on your storage budget. Petabytes can cost you a pretty penny if you're not careful

A8: #bigdatamgmt Beyond my pat advice (i.e., "plastics!"), I'd point to need NOT to treat #Hadoop as a "ha-dump" for never-explored data

A7 #bigdatamgmt well #dataintegration as we've discussed to ensure that acquisition isn't a bottleneck, also security, quality, governance
 Retweeted by jameskobielus

 Retweeted by jameskobielus

A7: #bigdatamgmt I'm going to add queryable archiving to key #bigdata success ingredients. Fetch older data you need when you need it

A7: #bigdatamgmt In terms of #dataintegration skills, you need SMEs who can prioritize #bigdata sources by downstream value

A7: #bigdatamgmt In terms of key #dataintegration ingredients for #bigdata success, I'd put strong data discovery/profiling near top

#BigDataMgmt A6: Data flow in #bigData? In 2012, it was more in one direction. #hadoop collected and distilled like refinery.
 Retweeted by jameskobielus

A6 #bigdatamgmt all ways - #Hadoop to EDW, EDW to Hadoop, Hadoop to Hadoop - more repositories = more integration
 Retweeted by jameskobielus

#BigDataMgmt A6: Data flow in #bigData? In 2014, data flow will be interconnected between many platforms/tech #hadoop #EDW #cloud #noSQL
 Retweeted by jameskobielus

A6: #bigdatamgmt You might have many #Hadoop clusters playing diverse roles in a multi-tier data env alongside EDW. Complex data flows.

A6: #bigdatamgmt If use #Hadoop only for landing/staging/ETL, one-way. If also use Hdp as sandbox & archive vis-a-vis EDW, multi-way

@mattbrandwein Yes. I plan to write about that next week.

A5 - and modern #dataintegration tools can push workloads to #hadoop for processing - e.g,.ELT #bigdatamgmt - complimentary, not competitive
 Retweeted by jameskobielus

@IBMbigdata A5 only if you consider having java or cobol 2 also be the same - kidding of course, #hadoop isn't #dataintegration #bigdatamgmt
 Retweeted by jameskobielus

A5: yes, using HDFS, Pig, MR properly in an ELT scenario to support a use case would be a data integration solution #bigdatamgmt
 Retweeted by jameskobielus

A5 - #hadoop distro persists #bigdata, #dataintegration is how it gets there. Complimentary, but clearly not the same! #bigdatamgmt
 Retweeted by jameskobielus

A5: #bigdatamgmt Having a #Hadoop distribution doesn't mean you have a low-latency data integ platform. #StreamComputing addresses that

A5: #bigdatamgmt Having a #Hadoop distro gives you fine platform for implementing some types of data integration (e.g unstr ETL)

A5: #bigdatamgmt Heavens to Betsy, no! Having a #Hadoop distro means you have a storage/analytic plat. But still need data integ plat

A4 #bigdatamgmt they must be integrated into existing apps/processes. best way to do so is leveraging existing #dataintegration technology
 Retweeted by jameskobielus

#bigdatamgmt A4: big data "has" to be integrated into the rest of the BI ecosystem because hadoop on its own can't deliver all the reqmts
 Retweeted by jameskobielus

#BigDataMgmt A4: #Big Data & #Analytics must be an Integrated Platform where ALL #Data types Integrate together seamlessly for success.
 Retweeted by jameskobielus

A4: #bigdatamgmt If each LOB implements its own #bigdata analytic platform, the siloes will be counterproductive for biz as whole

A4: #bigdatamgmt Tiered #bigdata arch of fit-to-purpose plats (RDBMS, #Hadoop, in-mem, #NoSQL, stream cmptg, etc) mitigates silo risk,

A4: #bigdatamgmt If deploy #bigdata tactically for limited uses, you risk siloes. If you deploy with architectural POV, mitigate that risk

A3: With #bigdata we are witnessing a transformation from transactional data capture to stream computing #bigdatamgmt
 Retweeted by jameskobielus

A3: #bigdatamgmt Downstream apps often connect #bigdata at rest & in-motion. Or #Hadoop platform may be connected with #streamcomputing

A3: #bigdatamgmt Many #bigdata projects involve "unstructured ETL," outputting structured data that may be connected to data from RDBMSs

A3: #bigdatamgmt Many #bigdata projects connect into unstructured sources--social, geospatial, etc.--to sentiment, experience analytics

#BigDataMgmt A2: #Data Discovery, Conditioning, etc. dominate the #DataScientist workday. Enhancement to Integration Stacks are critical.
 Retweeted by jameskobielus

A2: #bigdatamgmt Orgs spend an inordinate amount of time trying to find what they already possess. Especially older data.

well said! RT @dsrealtime01: @IBMbigdata #bigdatamgmt A2: ... risk losing it again in a #bigdata swamp if it isn't governed along the way.
 Retweeted by jameskobielus

A2 #bigdatamgmt Depends upon their needs..if it requires integration of tools to capture unstructured sets, it might take a lot of time.
 Retweeted by jameskobielus

A2 - far too little time, and they risk turning #hadoop into Ha-dump - integration & governance helps understand #bigdata - #bigdatamgmt
 Retweeted by jameskobielus

#BigDataMgmt A2: #bigData #dataIntegration soup? Tools will automate and productize just as #ETL did. (con't)
 Retweeted by jameskobielus

A2 - according 2 Forbes, 80% spent finding, integrating data. Don't handcode #bigdata integration- use tools and cut that time! #bigdatamgmt
 Retweeted by jameskobielus

A2: #bigdatamgmt Orgs with powerful data search/profling engines (e.g., Watson) can spend more time analyzing, less time finding

A2: #bigdatamgmt Orgs often spend upward of 75% of their time trying to find/integrate their data. Far less time trying to understand

@jameskobielus data integration should be a powerful part of the DB technology - not left to each developer to implement inefficiently.
 Retweeted by jameskobielus

A1: #bigdatamgmt Data integration is the middleware that makes it possible to manage #bigdata as a unified business resource.

A1:#bigdatamgmt Integration is the core part of accessing/capturing information. then use them for mining insights.
 Retweeted by jameskobielus

A1: in global corborations formed by merging independent companies data is spread all over the EIS and other data sources #BigDataMgmt
 Retweeted by jameskobielus

A1: #bigdatamgmt Without data integration, #bigdata can't be governed in any unified fashion.

A1: #bigdatamgmt Data integration can ensure that #bigdata you consolidate has been transformed, cleansed, & prepped for analysis

A1: #bigdatamgmt To ramp to "3 Vs" scale, may need to acquire data from disparate sources. Hence, #bigdata #dataintegration essential

@IBMbigdata I like to think of myself as an evangelanalyst. Or not.

@IBMbigdata @johnlmyers44 @dcorrigan A pleasure to be in/with such esteem company/companions today on #bigdatamgmt.

RT @IBMAnalytics: In < 2 hrs! #bigdatamgmt chat on “Spelling '#BigData Integration' from enterprise alphabet soup.” 

Join tweetchat "Data Warehousing on the Cloud" #IBMBLU Wed March 26, 11am-12noon ET 

Join tweetchat "Data Warehousing on the Cloud" #IBMBLU Wed March 26, 11am-12noon ET w/ACangiano, LKatsnelson, ARonthal, GHutchison, RAhuja

Privacy and #bigdata? Addressing the tricky contours of in-store privacy p1:  p2:  Wed #IBM q-h

RT @kexpplaylist #kexp Waking Light by Beck from Morning Phase JK--2014. Gr8 new soulful Beck LP. My fave of his since 2005's "Guero."

"Using Analytics to Detect Retail Fraud" ( ) JK--Discusses fraud management challenges & risk mitigation analytics

"Can #bigdata crunching help feed the world?" ( ) JK--Discusses data-generating IoT sensors that cattle have in guts

""Bayes Theorem & Monty Hall Problem" ( ) JK--Problem is finding someone younger than 40 with a clue who Monty Hall was

@TSManiscalco Funding source says nothing about addressable market. Different considerations.

"How Statisticians Could Help Find That Missing Plane" ( ) JK--Article from Nate Silver's new data-journalism site.

I've often wondered if football great Y.A. Tittle got tough kicking butts of dudes who made fun of his name. Like the boy named Sue.

"Nate Silver’s new FiveThirtyEight site launches but can his data-driven journalism find market?" ( ) JK--Niche at best

"Apache Spark-3 Real-World Use Cases" ( ) JK--Discusses experience of Yahoo and other early-adopters.

"#Hadoop Market Forecasted to Reach $50.2 Billion by 2020" ( ) JK--At 58 percent CAGR, says Allied Market Research.

"Please can we have a sensible discussion about "Big Data"?" ( ) JK--Absurd pleading wits-end tone of this makes me LOL

The terms "reactive" & "proactive" are useless. Everybody responds and everybody anticipates. Some connect the two faculties well. Some not.

I'm rather enjoying giving other people a platform for putting their thought leadership out for general consideration. #IBMDataMag

Anger is not becoming. Remember that curled up inside your animus is the declaration "i'm an anus."

I still love my Microsoft Zune, both the media player & the PC software. It's like Windows XP. Robust, trusty, & easy to use. Take that!

Social media can put you in continuous conversation with people from every phase of your life. It can be confusing. What phase are you now?

The next time I fly, I'm going to request an airliner that doesn't let anybody on board, including pilots, go anywhere near the transponder.

Let's review the strategic self-interest of US military assisting Malaysia in search for airliner in eastern hemisphere. Any thoughts?

TV news needs to introduce new slogan: "If it bleeds (and we have a video feed), it leads."

@TonyBaer I think there's an "open spruce" project on the Internet of Trees.

RT @MiaFarrow: Has TV 'news' gone completely crazy  JK--Apparently, missing airliner is Great White Whale of stories

"Curating Big Data Made Simple: Perspectives from Scientific Communities" ( ) JK--"Curating," as in "managing."

"Data Mining & Official Stats: Past, Present and Future" ( ) JK--"White-collar" v. "blue-collar" statisticians?

"Visualizing the Evolution of Rock Music" ( ) JK--Network graph & timeline of inter-artist influence patterns

Ah, root beer! Taste puts me back in summertime at an A&W stand somewhere in the north woods. Can't beat the Proustian sense remembrance!

"A Deep Dive into #IBM’s #BigData Strategy" ( ) JK--SVP speaks with eWeek ( ) #IBMDataMag

"Crossing the Data Chasm" ( ) JK--Video on challenges confronting chief data officers. #IBMDataMag

"Pushing the Limit" ( ) JK--#IBM Redpaper on DB2 very large database capabilities. #IBMDataMag

"Exploring the #NoSQL Family Tree" ( ) JK--Sam Bisbee provides useful infographic. #IBMDataMag

"What Makes In-Memory Computing So Next Generation?" ( ) JK--Nancy Hensley on #IBM DB2 w/ BLU Acceleration #IBMDataMag

"Going Beyond Data Science Toward an Analytics Ecosystem: Part 2" ( ) JK--Ahmed Fattah probing analysis in #IBMDataMag

"The Internet of Trees" ( ) JK--Steve Adler on "an arboreal Internet of Things" for global sustainability. #IBMDataMag

Drafted latest #IBM #AnalyzingMedia column: "The Irreducibly Human Center of Streaming Music Algorithmics."

I love how the words "analytics" & "algorithmics" sound together. Both have a liquid roll and a ticking click.

Internet of Things? Storage and protection of thing data over time ( ) Tuesday #IBM quick-hit

RT @kexpplaylist #kexp Cruel City by Augustines JK--2014. You'll be forgiven for thinking he's singing "come on a cruise ship."

#Datamation video roundtable, "How to Choose a #BigData Solution," that I did last week: 

Drafted new #Dataversity blog: "Gamification's Role in the Sourcing of Brand-Love Data"

Drafted new #IBM blog: "Plugging a lean, mean #bigdata analyzing machine into manufacturing"

A8: #cxo Businesses need to bring the real-time search experience into all their operational apps (OLTP, BI, etc.). Users need answers fast

A8: #cxo Businesses need to leverage the real-time data ingest, processing, and access features of existing #EDW.

A8: #cxo Biz succeed at real-time analytics by acquiring high-performance, low-latency, low-TCO data platforms.

A7: #cxo Real-time #analytics can play into either domain of decision management: decision support (human), decision automation (autom8d)

A7: #cxo Real-time #analytics leverages data in motion. Decision mgt derives insight from correlation of data in-motion & data at-rest

A7 Decision Management is muscle memory. Real Time Analytics is open eyes. #cxo
 Retweeted by jameskobielus

A7: #cxo Decision management requires alignment between decision cycles and data cycles. What needs real-time? What'll be fine with batch?

A7: #cxo Real-time #analytics ensures that every decision benefits from the freshest data/insights possible. Enhances decision mgt.

A6: The reality is all organizations have streaming data sources they can mine for insight. #cxo #analytics
 Retweeted by jameskobielus

A6: #cxo Any org that uses a search engine (their own or third party eg Google) knows full well that next-day response doesn't cut it

A6: #cxo Most orgs know they need near-real-time updates of key data, since many mundane operational matters demand immediate attention

A6: #cxo Real-time decisions are the stuff of every organization's daily life. Who do we talk to, work with, respond to AT THIS MOMENT?

a5. Real time decisions can be trusted when the right questions are being asked...asking the right questions is the hard part #CXO
 Retweeted by jameskobielus

A5: Some decisions req 100% confidence while others need 50% confidence to proceed #cxo
 Retweeted by jameskobielus

A5: #cxo R-T decis (by humans) are where BI & intuition proves value. R-T (by automated sys) thrives on pred analytics & biz rules

A5: #cxo Decisions "in the moment" are no more or less trustworthy than decisions "after long & careful consideration." Judge by outcomes

A4: #cxo In a real-time B2C #CXO environment, orgs must go deeper on data/analytic-driven decision automation & next best action

A4: #cxo In a real-time B2C engagement environment, orgs must tune their mobile apps to truly enable 24x7 access/updates/responses

A4: #cxo With real-time engagement, orgs must monitor whether B2C "alert overload" causing some customers to tune out & not-respond

A4: #cxo In a real-time env, orgs must tweak their customer-alerting intervals to support policy-driven levels of reminder & escalation

A4: #cxo In a customer-defined real-time environment, orgs need to provide self-service customer portal. Empower customers directly.

A3a: Engage w/ customers. Listen to their needs, wants and what works (doesn’t work) for them. #CXO
 Retweeted by jameskobielus

A3: #cxo Orgs can get customer time right by giving ability to self-svc/adjust infor/decision/response intervals.within engagement

A3: #cxo Orgs can get customer time right by IDing customers' decision/action cycles. Monthly, weekly, daily, continuous, adhoc?

A3: #cxo Orgs can get customer time right by developing customer engagemt "journeys" that phase/time each milestone from customer POV

A2: #cxo Right-time may resist real-time: customers may scream "uncle!" if you flood them with all info/updates at the moment they occur

A2: #cxo In engagement, right time is prescription driven by customer preference: when is best (to them) for biz to deliver/notify/etc

A2: #cxo Real-time is descriptive: awareness of event near-simul w/ occurrence. Right-time is prescriptive: may B OK to lag awareness

A1: "customer time" to me is amnt of time customer expects an answer. Could be immediate or days, depending on person & issue. #cxo
 Retweeted by jameskobielus

A1: #cxo If a customer's decision cycle is, say, weekly, and they only check updates on Fridays, real-time updates are unnecessary

A1: #cxo Real-time may be too-fast for customer-time, events flicking by too fast for customer attention/decision.

A1: #cxo "Real-time" means taking place live, now, in present time. "Customer-time" means taking place at customer-convenient time

Happy to giving you all this real-time experience, tweeting on #CXO real-time experience.

Between tweets, I shoveled the driveway. Now I'm back to shoveling the work tasks that are drifting inexorably up to my mind's awnings.

50 mins to #CXO chat: What’s real in real-time #custexp? w/ @s_v_g me et al  #nps #cxm #custserv #analytics

A shotgun secession vote doesn't constitute the legitimate voice of an electorate. Result invalid.

Recomm engines? Analytics grooving with whatever grooves groove you p1:  p2:  Mon #IBM qh #kexp

"#Bigdata: Personal analytics going beyond sports into realms of wealth & life satis" ( ) JK--Davenport on personal IoT

When I worked in downtown Detroit, oftn lunchd at Old Shillelagh. Crusty Irish barkeep annoyed cuz I always ordered Labatt's, never Guinness

"ML applied 2 affective computing (or how 2 teach machine 2 feel)" ( ) JK--With data from immersive multisensory envs

"How Many Data Scientists are out there?" ( ) JK--Gregory Piatetsky sifts thru th slippery stats on titles, hiring, etc

"#Hadoop Alternative Hydra Re-Spawns as Open Source" ( ) JK--Distrib task proc sys supporting streaming & batch ops

"Bayesian Inference – Made Simple" ( ) JK--Nice graphical illustration of difference from frequentist inference.

"Texas Sharpshooter Fallacy" ( ) JK--Well-meaning ways that data scientists falsely attribute patterns to random data

"Great example of root cause analysis" ( ) JK--Application of Monte Carlo statistical analysis 2 find missing airplanes

"The Ideal Data Science Team" ( ) JK--Good discussion enlivened by invocation of Christopher Columbus & Mark Zuckerberg

"Best kept secret re lin & log regress" ( ) JK--Granville: approach produces good-enuff approx sol'ns w/out stat libs

RT @kexpplaylist #kexp Cry to Me by Solomon Burke from Home in Your Heart JK--1962. Nice slice of gospel-inflected soul.

"#IBM Study: CMOs Fusing Int+Ext Data to Drive Financial Success" ( ) JK--Full study findings: 

Oh, Republican leaders are actually proposing ideas on healthcare insurance reform, not just ideological stonewalling.

Waiting 4 LinkedIn hack that asks "Does [friend X] know his/her a** from a hole in the ground?" All of mine definitely do. I'll vouch 4 that

RIP David Brenner. My favorite fill-in "Tonight Show" guest host during the Johnny Carson days. Big nose, nasal voice, snide quips.

I've been listening to Dick Vitale holler for 40+ years, ever since he was Univ of Detroit basketball coach. Amazing he still has a voice

Even rarer than the "man bites dog" story is the "dog eat dog." Canine cannibalism sells papers!!!!

Friend. If the person in question has any motive other than the pure desire to engage with you, they're not a friend. Just a sponge.

Annoyed that Malaysia Airlines getting a bad rap for this. We've flown it before & wouldn't hesitate to do so again. I recall good service.

We've suspended attending Sunday morning yoga class, due to teacher inadequacy. This morning, we did yoga ourselves in another room at club

That "it's gonna snow" smell again. I suspect it's 99 percent my own perspiration anticipating the shoveling I'll have to do.

Picked up bio of Apple designer Jony Ive. Put it down after umpteenth fawning reference to his "brilliance." Silicon Valley puffery.

WashPost used "decry" in 2 headlines today. This must stop! Need to employ a more representative cross-sampling of obscure English verbs

Something fishy about CBS Sunday Morning "Travolta's aging brain" interpretation of "Adele Dazeem." Perhaps he's dyslexic reading cue cards

I can't think of a better way to lose your sense of perspective on the day's events than to watch cable TV news.