Category Archives: Big Data

20
Sep

The Mystery of Just What Constitutes Big Data Continues

Posted by | Categories Big Data | Tagged , ,

Perhaps no other point so vexes me when I read about “Big Data” than the now familiar absence of a clear definition of the term. I just read an article published today, Thursday, September 5, 2013 in the “CIO” blog of the Online Wall Street Journal, Financial Services a ‘Real Leader’ in Leveraging Big Data. Michael Hickens, the author of this short post, makes a point about the new proclivity of financial services firms to adopt “big data”: “Banks and other financial services firms are further along than most other industries in making use of predictive analytics, according to a study reviewed by CIO Journal.”

In fact, this statement appears at the very start of Hickens’ article. But there is no connection made between “predictive analysis” and “big data”, so I’m left wondering about the point at hand, and where the author of this post would like to lead me. I’m also recollecting the late 1990s, when neural networks were, once again, a really strong area of interest of financial services firms. In fact, these businesses actively pursued the design of neural networks in an effort to advance the accuracy and utility of “predictive analysis”.

So what’s new about this time around? Beyond a mention of Hadoop as the data repository of choice, there is no mention whatsoever about the features I’m following on the topic of “big data”, meaning unstructured data, metadata tagging, etc. Are these financial services firms doing new work in these areas? Are they implementing taxonomies as a way of organizing unstructured data? Are they using metadata tagging techniques?

Unfortunately, Hickens short article does not include detail on these points. I would hope authors, going forward, try to be more specific about just what they mean by “big data” so reader like me can derive more benefit from articles on this kind of topic.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2013 All Rights Reserved

30
Aug

We See No Lessening of Market Interest in Big Data Despite Claims to the Contrary

Posted by | Categories Big Data | Tagged , ,

On Sunday, August 7, 2013, The online New York Times published an article in the “Opinion Pages” written by James Glanz, Is Big Data an Economic Big Dud?.

We couldn’t disagree more with Mr. Glanz’s opinion. A preponderance of information is now published online with markup languages, HTML and XML. This information is inherently unstructured, ASCII text data. Metadata tagging is an absolutely essential method of organizing this information into intelligible chunks for modern relational and non-relational database management systems (the most prominent application in the latter category is Hadoop from the Apache foundation).

Consumers want the lowest cost method of accessing all of this information. So web browsers function as universal clients. There isn’t a vendor of business intelligence (BI) applications courageous enough to buck this trend with a proprietary client. Big data technologies are not only the best method of addressing information in present form, they are, arguably, the only way to address modern chunks of information.

We wonder how an analyst could miss the point on this topic. Even small, smart mobile devices — smart phones and tablets — ship with browser clients. We can’t conceive of how it would be possible to roll back the years and revert to proprietary “application realms” (for example, Lotus Notes) requiring comparatively expensive consulting services from highly specialized personnel to extract meaningful results from the data stream.

Obviously, Big Data has magnetized some substantial amount of negative energy in the late summer of 2013. But there is no alternative, presently, to this technology as a method available to people looking to capture trends, etc, expeditiously, from literally tons and tons of unstructured information.

To argue markets are not interested in these trends, or other components of modern BI techniques, would be grossly inaccurate. There is more interest in exploiting all of the capabilities of BI than ever before.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2013 All Rights Reserved

11
Jun

On Recent Public Surveillance Controversies and the Unique Character of Big Data

Posted by | Categories Big Data | Tagged , , , ,

Until the recently publicized controversies over public surveillance operations in the United States, the real inescapable quality of Big Data eluded us. Until now, we had a hard time differentiating this new computing trend from earlier approaches to data management on a large scale. Massive databases are nothing new. U.S. Government agencies have maintained huge databases of information for years, as have their peers around the world. Very large businesses in the private sector have also maintained enormous data repositories of information. So what’s the big deal about Big Data?

From the controversies receiving a lot of attention from the press in early June, 2013, the real unique quality of Big Data have become clearer to us. The key was a recent article published on the CNN website Senate Intelligence Leaders Say Phone Surveillance is ‘Lawful’.

This article includes the following report: ” . . . [Senator Diane] Feinstein, D-California, said the government can only access the metadata, not the actual conversations that take place on the calls. After the information goes into a database, it can only be used if there is “reasonable and articulate suspicion that the records are relevant and related to terrorist activity.” So Big Data works with metadata tagging, terms and taxonomy characteristic of a large Content Management System (CMS) (for example, Microsoft® SharePoint®).

The CNN article unfortunately does not include a definition of metadata tagging for the general public. Neither is a definition included in an article on the same topic published in the Wall Street Journal (though the Journal article couples the term with the adjective “so-called”). Only the New York Times, in an article published on June 5, 2013, written by Charlie Savage and Edward Wyatt, U.S. Is Secretly Collecting Records of Verizon Calls includes a definition: “The collection of communications logs — or calling ‘metadata’ . . .”

But we disagree with the Times’ definition. In fact “metadata” amounts to information included on webpages within special tags, which are, in turn, used to present specific pieces of information (perhaps the content of a communication log, meaning telephone numbers, rather than a transcription of the conversation) for quick sorting (indexing) by webpage search tools. Bottom line, an unexpected benefit of the controversy in the U.S. in early June, 2013 over public surveillance by government agencies is promotion of one of the methods of Big Data, specifically the metadata tagging, terms and related taxonomic activity required to manage unstructured information found on webpages.

If your business can use a clearer communications method, please let us know. We spend a lot of time looking at how products, concepts and related components of computing are presented to markets and even the public. Please contact us to learn more.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2013 All Rights Reserved

22
Mar

Were Oracle’s Q3 Results a Harbinger of a Fundamental Shift in Enterprise Database Purchasing Trends?

Posted by | Categories Big Data | Tagged , ,

In article published in the March 22, 2013 edition of The Wall Street Journal, “New Rivals Clip Oracle’s Wings”, several wall street analysts present their opinions about the near term prospects for Oracle Corporation. We found several points of interest in the article.

Michael Hickins, who wrote the article with contribution by Rachel King, writes that “Safra Catz, Oracle’s co-president and chief cinancial officer blamed the sales miss on Oracle’s own miscues rather than underlying demand for its products.” But “Peter Goldmacher, who follows technology companies at investment manager Cowen and Co., said ‘there is a larger problem here’ than a one-quarter miscue. He noted that Oracle has had three disappointing quarters in the past two years . . . ‘new companies are offering equivalent or better functionality at a better price,’ he argued.”

Juxtaposing Catz’s comments with the quotes from Goldmacher lead us to doubt, at least the insight, if not the veracity of Catz’s comments from a skeptical position. But once we digested the whole article, we had to wonder if our reaction was not excessive. The drop in revenue only amounted to 2.3%, a miniscule difference for a large ISV with an annual revenue rate of $35Bil +. Why is this article leading us to this type of conclusion?

The answer is all about hyperbole. We attended the Bloomberg Big Data Conference last Thursday (seems like ages ago), on March 14, 2013. The observations on NoSQL “databases” (without rows or columns) in Mr. Hickens’ article, together with some of the conference commentary leads us to think that Wall Street analysts spending time tracking ISVs with Big Data solutions are on a hunt for the hackneyed “next big thing.”

But massive relational databases are very much like massive NoSQL “databases” (perhaps data libraries, or merely repositories would be a better term). The same type of tools are required to, once again, tag with metadata for browser, index and categorize based upon taxonomies.

We think the whole discussion is puffed up. Investors should exercise caution.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2013 All Rights Reserved

20
Mar

What is the Value of Big Data to the Healthcare Industry in the United States in 2013?

Posted by | Categories Big Data | Tagged , ,

The Bloomberg Big Data Conference panel discussion, “Rethinking Risks and Opportunities in Big Data: Healthcare” left us wondering whether the no SQL databases and related tools of big data can deliver high value to the healthcare industry in the United States in 2013. Ed Park, EVP and COO of athenahealth cautioned against simply implementing big data for implementation sake: “Pouring more money into the system hasn’t helped a lot for the last 20 years, and pouring more and more money in over the next twenty years is not necessarily going to help either”. We agree with his point. If IT systems fail to deliver meaningful value, then users cannot be expected to either keep using them, or look to add to them.

We got the familiar sense of a technology with questionable value from the healthcare panel discussion on big data. Opportunities to capture real descriptions of high value from the narratives presented by each of the panelists eluded us. If the best we can do is point to 23andme when citing an example of a business with a high value offer for the public, in our opinion something is missing.

On the other hand, we had a conversation with a colleague from the healthcare industry a year ago with something more meaningful to say on the topic. This individual held a technical management position in the Center for Disease Control (CDC) of the United States Federal Government in Atlanta, GA. The narrative he presented, for us, was much more compelling as he recounted how the Hollywood film “Contagion” actually depicted some of the systems in place at US CDC. The potential for massive repositories of data producing meaningful indicators of where threatening viruses are likely to spread next is a much more obvious demonstration of value than either purchasing one’s personal DNA profile, or crafting a highly persuasive “gentle” reminder to schedule a colonoscopy for someone turning 50.

The predictive capabilities of the US CDC big data systems are also an excellent example of a Cassandra whose warning is never ignored.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2013 All Rights Reserved