Analytics refers to the complex process of extracting insights from large data sets. Because IoT devices are taking data generation to new heights worldwide, IoT analytics is becoming a field all its own. Organizations across every sector are deploying connected devices to collect pertinent data, analyze it, and use the insights to guide decision-making.
Layers of IoT Analytics
Depending on the use case and goals of the IoT deployment, IoT analytics might look at different layers of information. For example, some industrial IoT sensors only collect data about the machines they’re attached to—with the goal of warning managers ahead of time if the equipment needs to be serviced. Other devices, such as water level sensors used in an agricultural deployment, are designed to collect information about their environment. Each method of leveraging IoT analytics looks at a different facet of data and utilizes it for a different purpose. Here’s a closer look at three possible approaches to IoT analytics:
Device Data for Predictive Maintenance
Many businesses use IoT sensors to monitor equipment in the field or on a factory floor. The sensors send data back to an IoT platform, where it’s fed into an algorithm. Typically, the platform allows managers to create triggers that will automatically send alerts if the sensors detect something unusual—if there’s a problem with a machine that indicates the need for service, for instance.
Similarly, organizations can use device data to monitor connectivity and address any issues that arise. For example, a micromobility vendor might be managing a thousand connected scooters within an urban area. They can use an IoT platform, such as Hologram Dashboard, to look at connectivity data and receive alerts if there are interruptions or anomalies.
Targeted Environmental Data
Many IoT use cases focus on collecting and analyzing environmental data. Sensors are designed to gather useful information about what’s around them—temperature sensors, weather monitors, security surveillance cameras, and telehealth devices all fall into this broad category. As sensors gather data, they typically send it to a cloud-based IoT platform, which may then connect to third-party analytics software or other enterprise systems.
Big Data to Enable Machine Learning
Big data refers to the massive data sets collected in the internet age, whether that’s consumer data gathered online or agricultural data gathered by sensors in the field. With IoT, there’s the potential to collect tremendous amounts of data today—far more than was possible in past decades. While some organizations choose to pre-sift their device data, looking only for certain cues that would notify them of a problem, others prefer to transmit all the data they can get, using it as fuel for powerful analytics and machine learning operations. They gather as much as information as possible and use it in many different ways.
For example, a healthcare technology company might choose to receive a constant stream of data from all of their telehealth monitors. After making the data anonymous to comply with healthcare privacy laws, they feed it all into a powerful AI-enabled software tool that can extract insights about device usage, patient outcomes, and more.
Let’s look at a few more examples of IoT analytics in action:
IIoT: Predictive Maintenance for Machines
The Industrial Internet of Things (IIoT) leverages the power of IoT to provide monitoring for large-scale production sites and other industries such as mining, construction, and fleet management.
A factory manager installs sensors on machines along the assembly line at a car factory. One of the machines starts to overheat, tripping an alert trigger in the management platform that warns the manager of an impending breakdown. The manager can then pause operations and repair the machine before the problem gets worse.
Smart Cities: Traffic and Parking Management
In smart cities, IoT sensors are placed in key places such as on traffic lights, in public transportation stations and vehicles, and on utility meters. City managers can then analyze the data sensors generate to isolate problems and drive better decision-making.
City managers install sensors in an urban parking garage. The sensors count the number of cars that enter and exit and display the number of available spaces on digital signage outside, making it easy for drivers to see if the garage is full. But all of that data the sensors are collecting is time-stamped, meaning that the city managers can analyze it to detect patterns and predict when the garage is most likely to fill up. This analysis can be used to make decisions, such as freeing up additional parking in the area during peak times or building another garage to accommodate more vehicles.
Smart Agriculture: Crop Care Decisions
Agriculture applications of IoT typically involve small, low-power sensors installed in remote areas. These sensors may only report back to the platform periodically, but over time the data yielded can paint a helpful picture that aids farmers in managing their crops and livestock.
For example, an agriculture enterprise installs IoT sensors throughout their fields that detect soil moisture levels. By monitoring these levels, not only can they make decisions in real-time (if moisture falls below a certain level, it triggers the irrigation system to turn on), they can utilize predictive analytics to foresee how much water they will need to nurture a particular crop from seedling to harvest.
Smart Home: Consumer Product Usage Analysis
Companies that produce internet-connected products can monitor how consumers are using them, creating a helpful swath of data that can guide future decisions about product improvements and eliminations.
For example, an appliance company monitors usage of its line of coffee makers using IoT analytics. They’re able to get a clear picture not only of how many customers are buying the products but also of how often they’re using them to brew a cup of joe. They can then layer that data with other information, such as mentions of their product on social media channels, to paint a fuller picture of how their coffee machine is faring.
Choosing the Right Connectivity to Enable Analytics
Depending on the type of analytics you want to perform, your devices will have different connectivity needs. If you’re streaming large quantities of data, you will need radio access technology that can support higher throughput—but if your deployment is more about triggered check-ins, low power solutions might be a better fit. A few possibilities for connectivity include:
Low Throughput: NB-IoT, Cat-M1 and Cat-1
Developed specifically for M2M/IoT use cases, these LTE standards can support many common IoT applications. Narrow-band IoT (NB-IoT) and Cat-M1 are low-power wide-area network (LPWAN) technologies created by 3GPP as part of Release 13 of the LTE standard. Both are low-cost solutions ideal for use cases such as health monitors, smart meters, and industrial monitors. Cat-M1 is faster, with upload and download speeds of 1 Mbps and a latency of 10 to 15 milliseconds—far lower than NB-IoT’s 1.6 to 10 seconds.
Cat-1 is designed for IoT devices with low and medium bandwidth needs. It’s an older technology and many carriers around the world have already adopted it. It’s more versatile than the other types, able to support higher bandwidth needs with much lower latency—but it also consumes more power and has a bit shorter signal range. Cat-1 use cases include asset tracking, point-of-sale terminals, retail kiosks, video surveillance, and some vehicle telematics data.
Higher Throughput: Cat-4 or Cat-6
These radio technologies are designed to support much higher throughput. Cat-4 has download speeds of 150 Mbps and upload speeds of 50 Mbps, while Cat-6 can support downloads up to 300 Mbps with a 50 Mbps upload speed. If your IoT use case involves streaming video, voice, or other data that requires precise real-time monitoring, one of these higher category radio technologies might be the best solution.
Simplifying Connectivity Analytics
While your use case might require a third-party application to analyze other layers of data, some IoT platforms provide baked-in connectivity analytics to help you manage your fleet. Hologram Inflight provides proactive connectivity monitoring, alerts, and collaboration features so you can track usage patterns, allowing you to monitor device health and location from afar. You’ll receive an instant alert when a device has potentially been tampered with or when fleet data usage rises above normal levels. These alerts appear in the Action Items menu in the Hologram Dashboard, where you can evaluate and manage incidents right away through. You can also route alerts to your organization’s Slack channel or through our Hologram API for quicker access.