The Pillars of Pangea
If you’ve ever met with one of our Pangea sales team, you’ll have heard of the four pillars of our portfolio:
Connectivity – Getting things talking.
Devices – Hardware if you need it.
Solutions – The whole end-to-end service wrap.
Analytics – huh?
Doesn’t it all end with solutions?
In the IoT world, we see that many deployments are often considered complete at the point of providing both devices and connectivity, or enabling an existing base of devices with connectivity. The fundamental task of getting the device talking to the cloud is completed, the application is receiving data, and all is well in the world with a fantastic IoT solution:
- The water sensor is reporting water levels to the monitoring system.
- The street lamps are turning on and off as expected.
- The tracking device is reporting its location.
- The toaster finally is talking to the fridge.
With a job well done, it’s tempting to walk away, but if we left at this point, we’d be walking away from some very useful data sets, often generated in real time, in front of our eyes. The art of analytics is all about capturing and interpreting these data sets.
Wiretapping data sets
If processing data is the end result (e.g. the application receiving the water sensor readout), then interpreting data is like plugging a wiretap into the data stream. By tapping into that data we can find new patterns and trends, and determine not only what’s happening right now, but answer why it’s happening, and what could happen the future.
In turn, all those insights can lead to more informed decision making, cost savings, operational efficiencies or even the creation of new products and revenue streams.
So, let’s go back to the example of the water sensor. Besides the application knowing when the water level is low, extra information—such as the rate of water level change between two time periods—could show statistics about drainage rates. Analysing these drainage rate statistics against another dataset, such as soil composition information from farms, or water piping routes in cities, would start to correlate regions where soil needs to be somehow managed, or where water pipes need to be changed due to leakages.
Getting these data sets together manually and enjoying the “eureka” moment can be fantastic as a one off. However, true value is derived when these data sets are processed in an intelligent and repeatable manner. Enter the analytics platform.
The IoT in many cases is about volume. Installing one smart water sensor, or one smart street lamp isn’t going to solve the customer’s—let alone the world’s—bigger problems. IoT solutions often have strength in numbers.
When we get one smart water sensor and deploy it across an entire region, for example, across a large farm or city, we start to get a steady flow of data points. Taking these data points and plotting them on a spreadsheet, our eyes may see certain patterns in numbers, but unless we’re in The Matrix we aren’t going to start visualising anything meaningful, and even if we did, that fleeting moment is likely gone.
So, if we as humans can’t achieve the task without great difficulty and time, we turn to the best brain in the business which loves nothing more than a repeatable task: the humble computer. These days there are a multitude of platforms all built to aggregate thousands or millions of devices and process all the data.
Getting data from an IoT enabled device is easier than ever. Publish and subscribe protocols such as MQTT are designed to aggregate IoT data, as opposed to one writing their own protocols, or spending unnecessary time coding a TCP/UDP stack, or re-using something such as SNMP to pull down data.
Message broker services are also allowing the solution builder to focus on the sensor data in the device by processing that data in the application via API. Managing connectivity to the wireless network, and even things such as IP addressing concerns or session management, are a thing of the past when using a publish-subscribe framework like MQTT.
There are many, many analytics platforms out there. Some focus more on presenting the data while others are more about complex AI and analysis.
A cool example
As a very simple demonstration about gathering data with Analytics, I’ve put together a Raspberry Pi, a DHT11 temperature sensor, and the Losant platform to graph the temperature and humidity of the Pangea office.
The Raspberry Pi gets a readout every few minutes, and publishes this data using the MQTT stack under specific “topics” of “temperature” and “humidity”. The Losant platform subscribes to these same topics, takes the data, and in this case plots it on a historical graph. Just as easily, it could be writing the values to a big table or a log file.
You can see this in real-time by clicking on the link here.
While in this case we’re showing the fundamentals of getting data, comparing other data sets could provide us with more valuable insights.
Stay tuned for my next blog post to read how we set up Losant to present this data, and be sure to check back to the Pangea website to see what the latest addition to our temperature sensor project is!
If you want to know more about Pangea, M2M or IoT, globally roaming SIM cards, customised solutions or analytics drop us a line at firstname.lastname@example.org.