Data generated by people and data generated by machines is actually quite different and as we move from the Internet of Things
to the Internet of Everything, this has some pretty interesting implications.
Data generated by things or machines is actually quite structured: A sensor is programmed or created to produce only a specific type of d
ata. Count the cars that cross the intersection, for example. And it’s predictable, sending a signal at specified intervals which makes the data pegged to a specific moment in time, as is the data’s relevance. It’s also generally low bandwidth, as you would imagine: A single signal from a sensor, providing specific data on a short time horizon.
Data generated by people, on the other hand, is highly unpredictable – I don’t know who I’m going to call or email and whether there’s a photo op when I step outside. Data from humans is unstructured, from spreadsheets to blooper videos, and has historical relevance. Tax returns, photos of your kids, the novel in draft in your desk drawer. It’s moderate to high bandwidth, depending on what you’re doing and it’s always on, always available.
So it’s important as we discuss the Internet of Everything to realize that we’re talking about different classes of data and that we’re now witnessing the birth of a new historical data set. Normally, big data and historical data increases in value with volume and time. The more you gather, over a longer period of time, the better it is. But what’s happening in this new world is that there’s a data-in-motion set, which is totally different. It’s location-based, with telemetry, motion sensors, traffic patterns, collision detection, for example. It’s real-time data. The data that tells you that you’re about to crash your car is not valuable if it’s stored, right? It’s most valuable when it’s available for immediate reaction!. So what you’re seeing is that the data value set is inverted from traditional data. Real-time data is normally gathered on a local basis, and then you have the option to store it or not. And to decide what part of the data to store, which might be just a couple of bits.
Understanding the difference is key. And then when you add predictive analytics, that’s where the wisdom starts coming in, that’s where the real breakthroughs happen.
This line of thinking can be extended yet further since a lot of data today is organised around individuals and from a static point of view.
Much better services (for people and for connected devices/machines) will become possible once data is applied dynamically i.e. time- and location-history as a basis for service offers.
Eventually, feedback loops will be applied employing “cause-and-effect” types of data.
For more on this topic and a view on how much personal data is worth see: http://www.more-with-mobile.com/2013/06/prices-and-value-of-consumer-data.html
Comments are closed.