American Elephants


Can You Trust “Big Data?” Not Really, Nor Little Data Either. by The Elephant's Child

big-data-skills-in-demand-are-a-mix-of-old-and-new-v21You have a computer by which you can visit this blog. So tell me, how do you get your news? The younger Millennials seem to get theirs from Facebook and Twitter and other social sites. Democrats rely on reliably Progressive websites, and Republicans assume from that bit of information that Democrats are unfamiliar with any websites that disagree with their conclusions.

Do you depend on “name” websites that you trust because everybody else seems to list them? Do you pay attention to a lot of individual bloggers (well, you’re here, so possibly you do.) What I’m getting at is who can you believe and who do you trust? What got me off on that theme was an article from February 22 of this year about “Big Data” in the Wall Street Journal, by Michael Malone. If the link doesn’t work, Google it, and do read the comments.

I wasn’t quite sure what “Big Data” was, compared to—little Data, so I looked it up. Merriam Webster: data: 1. factual information (as measurements or statistics) used as a basis for reasoning or calculation. 2. information output by a sensing device or organ that includes both useful and irrelevant or redundant information and must be processed to be meaningful. 3. information in numerical form that can be digitally transmitted or processed. That is, perhaps, helpful, but not exactly confidence building.

Hillary Clinton’s ‘Invisible Guiding Hand‘ had a statistician behind every strategic decision named Elan Kriegel.”To understand Kriegel’s role is to understand how Clinton has run her campaign—precise and efficient, meticulous and effective, and, yes, at times more mathematical than inspirational. Clinton advisers say almost no major decision is made…without first consulting Kriegel. ” That worked out well. But is perhaps a clue to Hillary’s uninspiring campaign.

At Maggie’s Farm, one of the group of authors had an article last year that I saved titled “Are We Overly Reliant on Data?” And his reflections on daring to ask the question.

A USAToday headline from September: The “VA quit sending performance data to national health quality site.” Saw an article today about a veteran who was unable to get the care he needed when a wound was full of maggots in a VA Hospital, and shortly died of sepsis.

From Climate Depot: “Italian meteorologist Colonel Paolo Ernan: Data manipulated to make people believe in global warming.” Well, yes.  It has long been apparent that alarm about global warming exists only in the computer programs devised to emulate the real climate of the Earth. They put into their programs what we know about climate, what we think we know, what they thought was likely and lots of pure guesswork. We know a little about El Nino and La Nina, for example, but we don’t know or understand much of anything about the actions of clouds. And if you want to know what is going to happen in 50 years, you’ll have to wait for 50 years to find out if you were right.

Holman Jenkins, writing in the Wall Street Journal at the end of August, 2014, “Big Data and Chicago’s Traffic-cam Scandal.”

Big data techniques are new in the world. It will take time to know how to feel about them and whether and how they should be legally corralled. For sheer inanity, though, there’s no beating a recent White House report quivering about the alleged menace of “digital redlining,” or the use of big-data marketing tactics in ways that supposedly disadvantage minority groups.

This alarm rests on an extravagant misunderstanding. Redlining was a crude method banks used to avoid losses in bad neighborhoods even at the cost of missing some profitable transactions—exactly the inefficiency big data is meant to improve upon. Failing to lure an eligible customer into a sale, after all, is hardly the goal of any business.

The real danger of the new technologies lies elsewhere, which the White House slightly touches upon in some of its fretting about police surveillance. The danger is microscopic regulation of our daily activities that we will invite on ourselves through the democratic process.

It seems that when you hear the term “The data tells us…” a caution flag should rise. You need to investigate a lot further. But everyone is relying on data, especially ‘big data.’ Hillary did, and is paying the price.  Her team pretty much shut Bill out. Bill certainly has some major problems with, um, integrity, but he has always had excellent political instincts. On the other hand, Hillary does not have any. But there you go, water under the bridge.

Who can you trust? Not much of anybody. Whatever it is, check it out. Governments at all levels are too ready to rely on what they are told is authoritative. We are all too dependent on our computers, but they are changing our world, and our dependence is making us more vulnerable.

ADDENDUM: Rereading this, I’m not at all sure I made myself clear. I am not railing at data. It is simply a fact of life, and as we use our computers, our choices and comments and what we just looked at becomes data. The search engines on which we rely for information— rely on us for information they can sell to marketers. If you drop by Amazon, as I did, find that they are having a sale on bras, you will be followed around the internet with a choice selection of what you looked at, all day.   As algorithms develop and refine searches, they will only get more intrusive. And we need good data, yet there is always the danger of over-dependence. Our only defense is a highly-developed common sense.

The case of computerized climate science is important. Climate alarmism, the belief that Earth is in danger, that the climate is changing and we have to save mankind has always been completely phony, but there are millions of true believers. The climate has been changing for millions of years. There are warm periods and ice ages. When the thermometers that the computer programs depend on for their data are situated next to air-conditioner vents or where acres of concrete reflect heat onto them, or they back up to a trash burner, the data is not going to be good. Ice cores help to tell about the past, tree rings are not so infallible.Satellite records are excellent, but don’t stretch far into the past.The rise of the oceans is measured in millimeters, not feet and in spite of Michael Mann’s claim that we’ll all be in trouble if the CO2 gets above 350 ppm,  greenhouses happily pump in 1,000 ppm to help their plants grow —you know there is something haywire about their data. Yet because the data is politically popular and governments act upon it, it becomes dangerous.


1 Comment so far
Leave a comment

As the author of the “Maggie’s Farm” article, I will add that in the 18 months which has passed since writing, nothing has changed. In fact the 2016 election shows just how more organizations have simply accepted data and become more reliant on it without questioning its value.

We continue to move further and further into the realm of ‘Big Data’ assuming this will solve all our problems without thinking about the additional problems this data is creating.

Like

Comment by Bulldog




Leave a comment