It usually starts with some ghoulish headline about “big data” and ends in a look toward the heavens, accompanied by a shake of the head and a deep, woeful sigh. The kind of big data they are talking about is not applicable to me, not yet, maybe not ever. It hasn’t reached the point where I can convince even myself that I could explain, at a kindergarten picnic, the difference between a structured and unstructured big data in an enterprise that has been collecting, and storing, “stuff” since before dirt and fire were invented.
And, worse still, It hasn’t had enough true, practical implementations to give me the ammo needed to convince any of the top brass what a “data set too large to be handled by traditional enterprises databases” would have the brass-neck to be doing in the confines of our shiny privately clouded data centers, let alone how throwing such an egregious data set on top of a smorgasbord of open source technologies with names as attractive as Pig and Zookeeper would suddenly reveal the path to true enlightenment.
The kind of big data that I am dreaming of doesn’t live inside Twitter, or Facebook, or enterprise file shares, or corporate email. It doesn’t care how many times the word “Coke” appears within the faceless ramblings of a marketeer’s unknowing congregation. No, the kind of big data I am dreaming of lives in the ceilings or behind the walls, in the air, underground, in every day equipment and (with a huge hat-tip to the venerable Dr Crosby) “the cloud in your pocket”.
Humans are (so writes my good friend Chris Hoff in a superb response to a really great post by Alex Williams) “an intelligently-complex species, and define even heady things like emotional responses as a function of two fundamental neurotransmitters — chemical messengers — the biogenic amines serotonin and dopamine” [sic]. The problem with humans, I believe, is that we are generally extremely inefficient as a species (compared to ants or bees) and, as a result of an eternity of cohabiting on top of this mantle of ultramafic rock that we call home, we have unwillingly allowed these emotional responses to become part of a bias when faced with making many kinds of decisions.
In both the posts referenced above, there are (probably correct) indications made that, over time, “machines will become social” and perhaps begin to form communities of like (excuse the pun) by using similar approaches to the ones we use today in our interactions with friends, family members, colleagues and so forth. Pretty interesting concepts, but, for the sake of my big data dream, I hope that machines never learn to display their emotional side.
Let me try to collate this a little.
I’ve written before about a universally understood metric named wrench time which, in our industry is a critical consideration for the owner/operators of many of the facilities that we build. It is so critical, in fact, that any delays to planned maintenance / outages caused by poor information and / or poor quality of work can realistically end up costing the owner/operators millions of dollars in lost revenue.
As buildings get smarter, the collective intelligence inside them grows exponentially and, as a simple example, the opportunities for “smart maintenance” get bigger and better.
The key to this is sensors. A colossal ad-hoc network of sensors. Yes, millions upon millions of them. They are already at work in many of the airports around the world, gathering vast swathes of information from emissions measurement and environmental compliance to security monitoring and advanced building management (for controlling lighting, power, etc).
Now add to that the hundreds of fixed installations inside any airport that need regular or break-fix maintenance, all with sensors of their own. Add location based services, socially aware machine to machine (cross-system to cross-system) communications, intelligent and proactive component refresh and you have the embryo of my big data dream.
But wait. There’s a problem. We still need humans. Of course we do.
In my vision for smart maintenance, we have to take one step further than our psyche may be programmed to take us. The machines know nothing of emotion. In the ideal world, the capabilities and historical performance of the humans that form part of the maintenance crew are known to the machines. The decisions on who should carry out the work (and of course, the tasks are issued to the worker via a web-socket aware tablet device which also tracks and reports back response times etc) that are based solely on the computed outcome of the combined data that the machines understand and process ahead of every request.
Then, a continual cycle (feedback loop) lets the machines learn, and adjust. They are free to share this information with other machines – they can build their own social environment where other machines are free to take and make decisions based on this information, but only ever based on facts. Pure facts, and always without emotion. Ruthless efficiency. Smart maintenance.
Of course, I would chose an aviation analogy, and not without reason. If you think this post touches a little on the “science fiction”, I can assure you this is very real. I’ve written this post from the point of view of the “back of house”, but what happens to all these sensors and the social interaction of machines when we find ourselves in the “self service airport” – will machines know and judge us by the data collected at the “front of house”?
You might not have to wait too long to find out.
Photo Credit : Peter & Maria Hoey from this WSJ Article on “The Self Service Airport”
- Big Data: you don’t need a data scientist, you need a data plumber (customerthink.com)
- Data Scientists: Meet Big Data’s Top Guns (informationweek.com)
- Reinventing Society in The Wake of Big Data (3quarksdaily.com)
- Of Airports and Thermostats (maneydigital.com)
- What Can Big Data Do for You? (arnoldit.com)
- When Big Data Questions Can’t Wait For Data Scientists (informationweek.com)