By Bill Meehan
Move over Smart Grid. A new buzz word has bumped you out of first place in the utility business. It’s Big Data. I’m not sure the term is even grammatically correct. Does it mean more data than you have ever had? Like the term Smart Grid, no one really knows for sure what it means. What we do know is that Big Data is big in the IT business.
I recently interviewed a candidate for a job and asked him what he knew about GIS. In his response he mentioned the term Big Data maybe 10 times and Smart Grid only 7. He didn’t say much about GIS. I concluded that Big Data was now bigger than Smart Grid.
So in the utility business, the biggest type of Big Data is all of the interval meter reads they get from the new smart meters, which is really not smart grid, but advanced metering infrastructure (AMI), which is not automatic metering (AMR) anymore, which the utility business bumped from being a buzz word several years ago. As you can see it’s not easy keeping up with these trends.
So in the old days of AMR, a utility read the meters once a month. Let’s say a single meter read consists of 50 bytes of data. That means a utility of 100,000 residential customers will need to store 60 mbytes (50 bytes times 12 months times 100,000 customers) of data for all the meter reads per year, not counting the commercial and industrial interval reads they have been capturing for years.
Now, for 15 minute interval AMI reads, assuming the same 50 bytes per read, the company will need to store somewhat less than 200 gbyte (50 bytes times 4 times per hour times 24 hours, times 360 days in the year times 100,000 customers) of data for all its AMI data each year. That’s a couple of flash drives worth of data. My laptop computer has a 150 gbyte drive. So does this qualify as Big Data? Sure, but I’m oversimplifying the situation. There are lots of other data being read by AMI.
I believe the buzz around Big Data has more to do with what you do with the data than the actual volume of data. So it’s really the context for the data that is critical. If a customer wants to reduce consumption and save money, it’s more important to know what’s running during a high consumption interval than to have 15-minute granularity of consumption data. Maybe when we get to dynamic pricing this will change.
It seems to me it’s all about context.
Big Data plays a strong role in politics. An incumbent Democratic president, running for a second term during a time of huge economic turmoil and high unemployment (inherited from a Republican president) facing a wealthy Republican governor serves as a reminder that Big Data is not about volume, but about context. Of course that was the presidential election of 1936.
In 1936, the publication Literary Digest sent a survey to 10 million voters asking who they were going to vote for. Nearly a quarter responded. The results strongly indicated a win for Alf Landon, the Republican. Roosevelt won in the biggest landslide in history up to that point. Since then pollsters such as Gallop and others learned that sheer volume of data doesn’t predict winners and losers. Context matters. If Literary Digest had asked a few context questions, they would have discovered that the people who returned the survey tended to vote Republican anyway.
This was Big Bad Data.
By now I am sure you are asking, “What does all this have to do with spatial information and GIS?” Even in politics, location matters. In fact it matters a lot. If a utility wants to know how to use their AMI data to drive better business decisions, it needs to look at the context of all this consumption data. What’s the relationship of the consumption to demographics; of shifting weather patterns, not just temperatures, but shifting populations? What will be the impact of certain patterns of consumption on the low voltage wires and transformers feeding neighborhoods? Where will the adoption of electric vehicles have the most negative impact? How will utilities assess changes in pricing policies? Who will be the winners and losers? More importantly, where will they be and what will they be thinking and doing? Big Data will be a part of that. All of that context is about location.
Really Big Data is probably coming from social networking. But even that Big Data can be misleading if not put in context. If a utility just looked at tweets from customers for negative comments, they might get just one impression. Not every segment of the population tweets. Utilities can mine that data to figure out just what people, and what kinds of people, are thinking about and then integrate that data with their GIS. They can associate the negative tweets with spatially enabled customer satisfaction surveys, with places were outages have been frequent, with places were construction activity has inconvenienced customers, or where the utility has cut down a favorite tree.
Big Data without context may lead to placing bad bets.
There is no doubt that AMI data is pretty big. But it has a very narrow focus without some additional context. The challenge for utilities will be to avoid using Big Data (even if it is not that big) in isolation, but to combine it with all kinds of data to put it all in context. One of the most powerful tools for determining spatial context is GIS.
