Recently, Gartner published a research note about the way IoT data will quickly dwarf social media data. I agree and have been predicting this phenomenon for quite some time. Gartner’s findings are in line with the way I see growth happening, but there’s an additional point I want to emphasize: It’s not just the size of IoT data that will dwarf social media data, but also its value.
IoT data has the ability to create an ultra high definition view of your landscape. You’re going to get the bits and you need to store them. That’s not the story. What you need are the resources to assemble those pixels into a coherent picture.
Something like a digital photo with remarkable detail that can zoom out for the equivalent of a majestic-earth-in-space holistic view and then zero in on matching grains of sand on beaches around the globe.
This is where another perspective on IoT comes into play that isn’t talked about nearly enough: the Analytics of Everything. Now is the time to prepare your analytics capability to effectively utilize IoT data. This requires thinking not only about the data collection environment, but also aligning the technology and human skills needed to conduct advanced analytics on that data.
So in terms of building up your analytic muscle for the IoT, where should you start? IoT data has a few common elements that suggest where you might start to bolster your analytics capability.
Nearly all IoT data has a significant geospatial component; location is almost always part of the data produced, even if it is associated with other detail. Sensors are also making their observations over time, which points to temporal data management.
Time series and path analysis will play a very significant role in making use of IoT data. These two areas, spatial and temporal analytics, are fundamental to the IoT, yet few companies have invested in their capability to make use of them, let alone support them at the scale of the data about to be encountered.
What types of expertise will you need specifically? You’ll need data scientists with a background in spatial and temporal analytics. Data scientists explore the data and come up with questions that IoT data can answer. Data scientists develop data products, which will ultimately be consumed by BI analysts. Another key, but less frequently mentioned, role is data engineer. Data engineers have an IT background and can take the data scientists’ raw data product and clean, normalize, and model the data so that it can be consumed by BI tools such as Tableau or MicroStrategy.
You need data scientists (who envision the questions) as well as data architects (who get the data ready from an IT perspective) to form a pipeline for providing IoT data to an army of BI analysts, who can then answer a myriad of business questions. Make sure you have at least some BI analysts with spatial and temporal analytics skills.
When it becomes possible to measure almost everything, whether it’s people, fixed infrastructure, or a poker chip in a casino, every industry and business is liable to develop its own, unique use cases. Individual bits of IoT data have inherently low value density, which is why their value only becomes apparent when their quantity is massive. With that volume comes a corresponding need to filter data at scale and “promote” the detail that delivers high value into a data product while filtering out the noise that doesn’t matter. Along with storage, you need greater smarts about the data, working toward an ecosystem capable of delivering the right data to the right analytic.
Analytic expertise needs to be developed in-house in parallel with IoT deployment. If you’ve already primed the IoT pump but don’t have the corresponding analytics capability in place, the time to acquire expertise is now. And when you bring in an IoT consultant, make sure someone is shadowing her. If the IoT is strategically important to your business, you need to own some of that intelligence and develop analytic capabilities to take advantage of IoT data and explore use cases.
Ultimately, the Analytics of Everything is the measure of how smart your business is about the IoT and how much value you can derive. These two metrics are much more important than the size of your data.
No comments:
Post a Comment