How Space-Time Makes Sense of Big Data
As the number of people connected to the web continues to grow, so too does the vast amount of information about those individuals. Putting this massive data store to use improves predictions and improves the overall quality of data but can also increase computational speeds. What happens when you add mobile devices and locations to the mix? That geospatial data attribute feeds big data analytics like a super-food and creates “space-time travel” data.
At GigaOM’s Structure Big Data conference on Wednesday, Jeff Jonas, a IBM Distinguished Engineer explained the concept with some startling examples. With roughly 600 billion data transactions from cellular phones on a daily basis, adding space and time to traditional data objects can help predict where someone will be on a given day and time with up to 87 percent accuracy, for example. Adding space-time works because, oddly enough, of physics, says Jonas.
Expert counting is a traditional way to observe data, but the same thing cannot be in two places at once. Including space and time observations removes ambiguity. For example, the last 10 years of address history, taken in context, can tell if a person is the same or not, when digging through billions of rows of data.
Why would companies want to add the physics of space-time to their data efforts? More context is needed, Jonas says, because the amount of captured data is rising faster than the creation of algorithms to make sense of the data. That’s creating a gap in the understanding of vast information stores and adding space-time adds much more context. Observations are where the sense-making begins but without context, it’s like trying to build a jigsaw puzzle without the picture of the final product.