point

Contextual Analytics on IoT Data with Micro-location

Abhishek Singh, Co-Founder & CTO, Evoxyz Technologies
Monday, February 13, 2017
Abhishek Singh, Co-Founder & CTO, Evoxyz Technologies
Headquarter in Gurgaon, Evoxyz Technologies is dedicated in proffering IoT services and aims to create a safer world for children by leveraging technology. The entity develops innovative child safety solutions that enable all parents to get safety alerts and information about child whereabouts.

IoT Analytics

We are lucky to be living in a world where everything is supposed to be smart. I live in a smart home, I know the time with a smart watch, I talk and manage my life on a smart phone, my car is a smart car and soon my city would be a smart city. The sea of these smart devices talking to each other and to the cloud everywhere constitutes what is termed as IOT. With so much of supposed smartness going around, there is a need to understand the holistic view of these smart devices together instead of looking at them as each single device and using their data to enrich our lives. The analysis of combined data generated by these devices, if properly interpreted, could be much more than just the sum of their individual analysis. Hence, we need IOT analytics. The need to study this data together, as it happens, in near real time, to know what happened and why it happened so that it could be predicted that what might happen and corrective actions be taken automatically, if needed.

The Relevance of Context in Analytics

Analytics is a statistical and mathematical tool to study data. Any insights generated by analytics can be only as intelligent as the data which it works on. There are multiple ways to do analytics. The classical way is to hypothesize a pattern model, collect data to prove that trend and then observe how that trend behaves in the future. Similar models could be made for anomalies too. Newer approach with analytics however does not work on hypothesis. Multiple data sources are collated and associative analysis is done to mine out hidden relationships which cannot be seen by naked eye or thought of before-hand. This leads to new hidden insights in data which were otherwise impossible to gain with classical methods. As we gather multiple data sources, what comes to light is the context of the trend or the anomaly. As an example, if we look at a five day data of a telecom provider, a surge in SMS usage on its own just tells us that it is an anomaly. When we club it with a date, we could get to know that this is because of a festive occasion or as we club it with historical data, we might find out that it is a fraudulent pattern or as we just gather more data, we might find out that it is just a periodic anomaly which comes every month and hence not an anomaly at all.


Share on Twitter
Share on LinkedIn
Share on facebook