Wednesday 27 April 2016

Big Data and Bakasur

It is seven thirty in the evening. I have had a tough day, extremely busy, no time for lunch. Five more minutes, before I complete the task on hand, and head out for dinner. That is when I hear a big noise and go out on the balcony to investigate. I don’t find anything and as I look at where the 3X optical zoom vision, limited by the buildings in the horizon, takes me I find a car parked across the street with its hazard lights on.

Not a sight, which should make you think twice. Just when I am about to head back, a piece of data from the brain sends information that my car was parked in the same place. I look back and instantly recognize that it is my car, which has these hazard lights on. These are times when you realize that the brain is a supercomputer. Its ability to retrieve information from different segments and piece it together to give the analysis, result and repercussions can give any advanced analytics program running on high speed processing machines a run for its money. In simple English it is called intuition
.
I suddenly realize that the hazard lights have been blinking from noon; for seven hours. Reality strikes that the battery of the car would have drained and it will not start. I run down the stairs, cross the busy road in James Bond style, getting a few cars to screech and shout at me. I open the car, enter it, switch off the hazard lights, and hoping against hope insert the key into the ignition. As expected the car doesn’t start.

I call the car’s on road toll free assistance number and am promptly greeted by a tele-caller. My wife says men are like kids. When kids bawl, mom knows that they are hungry. When men get angry, the major reason is hunger, the incident confronting them is just the trigger. So if anybody of you has an irate boss, check his food schedule, chances are he overworks and skips food. So my brain was at a flashpoint because of the hunger and anger at my stupidity. It was no surprise that it exploded on the tele-caller.

Before I say anything about the process, I will commend him, and say that he did his job very well. 
Kudos to him, for tolerating an irate customer. However I did an analysis of my call. My call ran for sixteen minutes out of which I spent last three minutes on berating the tele-calling process, not more than two in telling my problem and my location and spent more than ten minutes either on hold or providing a host of data.

The guy asked me my details about my car and then sent in a bouncer. “Sir can I have the chassis number?” I started wondering, how did he know I was working with Chase and gave him my employee number. “That is incorrect sir, can I have the chassis number.” In the next five minutes I started going through the documents of the car hunting for the number which would decide whether I would go home in four wheels or three. It was nearly eight at night: I found it and tried reading it with my blurred eyesight (I am forty plus) only to find that it is a forty eight digit alpha-numeric number. Sorry maybe it is fifteen. All numbers beyond six digits are the same to me. Come to think of it, there can be no safer password than the chassis number of your car.

Finally he gave up on me and retrieved it from customer care. I was told that a complaint can’t be registered without the chassis number. Then he went to ask me so many other questions, before he asked me my problem and location, I broke into a sweat. I seriously hoped he would not ask for my grandfather’s PAN number. Thankfully he didn’t.

I then began wondering, why did they need all this information. The answer -  data and analytics. This information would one day find its way into my insurance records. Does this guy leave his lights on, and hence, how often do I have to provide him road side assistance? At what mileage did the problem happen? What color is the car?

It is the age of data, Big Data as they call it. There is an entire industry out there which thrives on analytics and compressing big data. Sometimes I wonder, whether we need to massage data on 30 parameters or 5 to get ‘accurate’ results. Will there be a substantial difference in results by increasing the number of parameters and different types of analysis? Or is advanced Analytics a mirage by itself?

From a customer service perspective, I only cared about two things:
  1. Have they recorded my problem correctly?
  2. Have they got my location right


I am sorry, but as a customer in a problem I do not appreciate the car company’s need to collect data for analysis. Not all functions can be a source of data collection, and someone needs to just look at what that function is supposed to achieve.

Accuracy has a cost. The more variables, parameters you analyse, more data you collect. Sometimes I think, there isn't much difference between‘B’ for ‘Big Data’, and ‘B’ for ‘Bakasur’, a monster, with a great appetite. The villagers had to provide him huge amounts of food and one member of the family, whom he would eat. ‘Big Data’ and ‘Analytics, have a similar voracious appetite. I hope it does not 
end up devouring the users.

I am writing this blog sitting in the car, as I have to charge the battery for a minimum of an hour. Time to refuel.