Big Data: Some Obvious Issues

November 30, 2015

Imagine this. Information Week writes an article which does not mix up marketing jargon, cheerleading, and wild generalizations. No. It’s true. Es verdad.

Navigate to “Big Data & The Law Of Diminishing Returns.” The write up is a recycling of comments from and Ivory Tower type at Harvard University. Not enough real world experience? Never fear, the poobah is Cathy O’Neil, who worked in the private sector.

Here are her observations as presented by the estimable Information Week real journalists. Note: my observation appears in italics after the words of wisdom.

“The [Big Data] technology is encouraging people to use algorithms they don’t understand.” My question: How many information professionals got an A in math?

“Know what you don’t know. It’s hard.” My question: How does not know oneself if the self is trying to hit one’s numbers and work with concepts about which their information is skewed by Google boosted assumptions about one’s intelligence?

The write up includes this bummer of a statement to the point and click analytics wizards:

“I’d rather have five orthogonal modest data sets than one ginormous data set along a single axis…That is where the law of diminishing returns kicks in.” This is attributed to Caribou Hoenig, a venture capital firm. My question: What is ginormous?

The write up also reveals, without much in the way of questioning the analytic method, that IDC has calculated that the size of the Big Data market will be “$58.6 billion by the end of the year, and it would grow to $101.9 billion by 2019.”

Perhaps clear thinking about data begins with some thinking about where numbers come from, the validity of the data set, and the methods used to figure out the future.

Oh, right. That’s the point of the article. Too bad the write up ignores its own advice. I like that ginormous number in 2019. Yep, clear thinking about data abounds.

Stephen E Arnold, November 30, 2015

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta