Analytics of Things: The Beginners Guide.
The Analytics of Things (AoT) makes possible to understand the Internet of Things (IoT), and apparently the infinite network of everyday objects that share information and execute tasks in conjunction with other objects. In order to convert these connected “things” into smart objects, analytics are necessary for those things to take concrete actions based on certain triggers.
Let’s describe a simple example: I need to leave in a hurry because the supermarket is about to close and need some more bananas for breakfast. But, I’m a scatterbrain and forgot to turn off the oven and the gas burner. As I leave my house, my GPS location is sent to the cloud where it is analyzed. The algorithms determine that I am moving quite fast (I’m in my car) and that some of my household devices are inadvertently on. In this particular case an off signal is sent to the oven’s and the gas burner’s kill switches, reducing unnecessary risk. Other actions could take place after placing a confirmation request via instant message (which of course would be read to me with my mobile device’s personal assistant, since I’m driving), such as locking all doors, shutting the lights off, etc. This could even be accepted by an insurance company to reduce my homeowner’s insurance policy since I’m better protected from unnecessary risk.
Other examples include: public trash cans that compact waste as needed and signals city workers when it’s full and ready to be cleaned, sprinklers that detect rain and automatically shut off…the possibilities are truly unlimited. All of this will be possible with the IoT and AoT.
The IoT consists of three main components:
- The “smart” things or objects themselves
- The networks that connect them
- The systems that collect and use the data flowing to and from “smart” things
In order to improve our lives and businesses, this data has to be analyzed to be useful. It’s a fact that analytics are necessary to make connected devices smart. So, AoT intervenes in the last element above to help us use the full potential of the IoT.
One of the main virtues of analytics that help connect smart devices is that data can be amassed from multiple sources and then compared, leading to better decisions (much like human-decision-making with a punch). This is where the data scientist comes into play. Data scientists have the interesting task for training algorithms in order to deal with this data and these events. So an algorithm would provide actions based on specific results such as:
If (eventRisk == “red”) then send kill signal to device A, else (eventRisk == “amber”) then send alert signal to ownerCellPhone.
However determining how something is determined to be “red” or “amber” may not be so simple. Regression analysis? Instance based methods? Decision tree learning? Bayesian methods? The options are many, but the algorithms need to be trained, tested and sent to production for real time analytics to work.
We’ll learn much more about the Analytics of Things as these connected devices thrive and become more common in our lives. We’ll also learn what kinds of analytical functions are most helpful for us. For now, it’s good to remember that The Internet of Things is only useful if those said things are smart, and that will only happen with the help of all good data scientists out there and through The Analytics of Things.
Would you like a demo for 3Blades? Enroll for one by clicking on the button below!