A Framework for IoT and Data Analytics

At the heart of deriving value from an IoT implementation is not just data – it is data analytics. IoT “listening posts,” such as devices and sensors, are throwing off a lot of streaming data, but to make sense of it and take meaningful downstream actions requires driving the analytics process efficiently and at scale.

A framework to think about harnessing this data requires one to think about two dimensions: first, understanding the types of data that can be analyzed; and second, how the end-user will use this data. Data can be very different in nature. Normally one would think of the data that comes straight from the sensor that being used to perform various kinds of business enablement use cases. However, there is also metadata that can predict device behavior, anomalies and security issues. In addition, your network provider may provide to you usage data – bytes/dollars in a given period.

Looking at the second dimension of how an end-user would engage with the data, there are five essential steps an end-user should follow to build a successful data path:

• Establish goals
• Gain visibility
• Identify opportunities
• Perform actions
• Track outcomes

If you bring both of these dimensions together, the two-dimensional framework allows one to think clearly about a variety of use cases, depending on the type of data and the end user in mind. If you are in the fleet management industry, where KORE has deep expertise, here are three “free” examples to illustrate the approach:

Example One:

• Type of Data: Telematics sensor data.
• User: Fleet manager.
• Establish goals: Optimize fuel utilization across fleet.
• Gain visibility: View start, stop, and idle times across the fleet.
• Identify opportunities: Pinpoint specific fleet outliers around start/stop times and create targets around optimal engine start-stop rules.
• Perform actions: Make recommendations for driver protocols and future vehicle purchases based on data.
• Track outcomes: View start, stop, and idle time trends and fuel consumption trends.

Example Two:

• Type of data: Metadata regarding authentication and session.
• User: Security Director.
• Establish goals: Spot security issues with devices (anomaly detection).
• Gain visibility: Observe device behavior by auditing metadata such as flow of traffic, times of connection, etc.
• Identify opportunities: View changes in device behavior against a “profiled” behavior to take appropriate action. E.g., the normal trend is to send 100KB per day, but since yesterday, the device is sending 30MB (is someone is starting to use the device SIM in their iPad?).
• Perform action: Create policies that address anomalous device usage that raises security threat levels.
• Track outcomes: View device behavior anomaly trends.

Example Three:

• Type of data: Cellular usage data
• User: Finance team member looking at total network costs in the IoT Fleet program
• Establish goals: Create optimal rate plan allocation.
• Gain visibility: View connection inventory and associated rate plans, month-to-month usage across account, and usage for individual assets.
• Identify opportunities: Identify assets consistently resulting in overages, asset outliers with respect to similar assets and opportunities to switch off, and assets that incur account access fees and no usage.
• Perform action: Initiate audit of all areas that result in higher usage costs.
• Track outcomes: View usage and billing trends along with usage outliers.

As IoT deployments scale up in organizations from a few hundred devices to hundreds of thousands of devices, putting a use case mapped against the type of data used and the five-step process outlined in the framework above, can help realize the different types of value IoT data can unlock.

Learn how a partnership with KORE will can simplify the challenge of data analytics to help meet your IoT goals.