Is your analytics strategy ready for the IoT data deluge?
“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.”
― Arthur Conan Doyle
A recent report by Gartner shows that 50% of the businesses would integrate IoT systems in
some form or manner, by the year 2020. This means, businesses would have access to a lot
more data, at an unprecedentedly faster clip!
And it is not just about scale. The variety of data being generated by the integrated IoT
devices and sensors, mean that, traditional analytics approaches, need a relook.
Talking from an analytics perspective, here are a few things that make it a lot more
interesting and challenging:
- To consume the volume of data being generated, new techniques are required to
capture and interpret data.
- In most cases, to analyze this huge amount and variety of data, it becomes necessary
to distribute the analytics over multiple devices servers; combining the results, to
paint a coherent picture. For example, in some cases, a simple query at a device level
can simplify things, instead of aggregating at a server level.
- The expectations from a business perspective is to build real-time analytics
capabilities. Or, perhaps, going a step further to predict potential problems and/or
- Although highly aspirational, automating complex scenarios including mission critical
systems, could be the next logical step. For cases where sensitive systems are
affected, security and reliability are the key factors.
- Building real time capabilities does not mean doing away with persisting data for a
long duration, for batch analytics. This is essential, for use cases like finding trends
Having said that, all technology decisions, eventually boil down to the problem we want to
solve! Here are some of the important questions we need to ask, before deciding on a
- How ‘real-time’ should the analytics be? Should we include advanced capabilities like
- Should we represent the outcomes in the form of information & insights, or, go a
step further to automate the processes based on the analytics?
- What kind of data should we collect, and, what is the useful bit of information we
want to wean out of it? For example, a lot of sensors and devices collect images, but
facial recognition vs. a security logging mechanism, are two entirely different use
- How long should we persist the data for historical reference, and, should we store
raw data or store aggregated results? For example, analysing seasonality trend over
different quarters might not need data stored per millisecond!
This is, by no means, an exhaustive list; but I hope, it is indicative of the challenges and
opportunities that come with increasing IoT integration. Asking these questions and
answering them in the specific context of your business becomes crucial in framing the right
analytics strategy and adopting the right bit of technology. And that’s precisely what we
have been doing at BluePi, bringing people and innovation together, to solve complex
business problems for our clients!
This blog has been written by Mr. Gaurav Batra, one of our core Leadership Team’s Member
Gaurav Batra, Head, Products
With over 12 years of experience across multiple roles in client delivery and app development; Gaurav developed a stronghold in analysing the requirements of legacy applications, before eventually switching over to database design & development and taking a fancy to the cloud.
Having completed Masters in Computer Applications (MCA) in 2004 from Institute of Management and Technology, Faridabad; he’s been associated with leading brands like Hewitt Associates, ACS (A Xerox Company), and Cognizant, before joining BluePi in 2014.
Under his leadership, BluePi has built a suite of custom products designed exclusively for Media & Entertainment businesses. The suite, piStats, takes care of all digital needs of a contemporary media entertainment business. From a custom CMS, to immersive user experience via mobile & web applications, tools to manage digital assets, analytics and recommendations – piStats does it all.