4 Useful Applications of Predictive Analytics on Streaming IoT Data


Internet of Things (IoT) data is expanding at a tremendous pace. Any IoT stream has too much data for a human to analyze in real time. For most companies taking in IoT data, few applications even present historical IoT data in formats that are actionable by people. Still, companies need to make some use of IoT data. Otherwise, IoT data is stored just because it exists. Without application and a business purpose, the IoT data is a cost, not an asset.

The problem of having too much data for a person to work with can be overcome through machine learning and predictive analytics. Traditional business intelligence only shows what happened in the IoT data stream in the past. Predictive analytics tells people and enterprise systems what to do now.

Process Improvement

Predictive analytics (and related concepts such as artificial intelligence and machine learning) use historical data to create a model. The model is statistically based, using algorithms that examine the accumulated and stored IoT data. Think of the model as a set of rules the algorithm learns from the historical data.

The deployed model looks at the streaming IoT data in real time. It applies the learned rules to identify actionable circumstances or trends. Once identified, the system makes predictions about near-term, future events.

While the potential use of IoT data is wide ranging, here are four common useful applications of predictive analytics on streaming IoT data.

1. Predictive Maintenance

Probably the most widely advertised use of IoT data for business is regarding maintenance and seeing when machines or systems need to be worked on to prevent problems. A commercial features an elevator technician showing up at the office building to perform work on an elevator before it breaks. This is predictive analytics over IoT data.

The straightforward logic is if a particular part or sensor is out of its normal operating range, then dispatch a technician. People get that idea. Predictive analytics takes it a step further by examining the trends of those same sensors to predict when it will exceed normal ranges.

2. Automatic Stocking

Using IoT data coming from products that a customer uses lets suppliers keep track of their customers. Assuming the customer allows the supplier access to how and when items are being used or sold, the supplier can track usage and predict when to resupply.

The benefit to customers is they will be well supplied. Suppliers proactively keep their customers stocked, which results in less cost to retain and service them. Predicting when to ship products automatically reduces operating costs and ensures a more stable supply.

3. Predictive Visits—Customer Acknowledgement

Retail companies would love to know who their representatives are talking to when they enter the store. IoT data lets the retailer or industrial supplier know when a customer is about to enter the store or location.

By applying geofencing to an app on a smartphone or customer device (think IoT location data inside a product), the business can see when a customer is getting close or entering the business. Predictive systems then look at the customer’s history and profile to present the best product to cross-sell or most likely reason the customer is coming by.

One consideration is to keep the predicted information focused. Presenting too much information to act upon in the presence of the customer often leads to confusion and inaction.

4. Supply Chain Management

As many items move within the supply chain from manufacturer to ultimate user, tracking location and status using IoT allows multiple levels of the supply chain to work together. Sharing IoT data in the supply chain enables the integration of data to create better analysis and predictive models.

Imagine a cooperative arrangement where the ultimate customer’s use of an item as shown by IoT data is shared back up the supply chain. Suppliers and vendors up the supply chain can aggregate data from multiple customers below them to create much better predictions on future needs.

Those same upstream suppliers can share their IoT data to their customers to inform them of where orders are, lead times, quality, or other potentially useful data elements. The downstream supply chain company can use that IoT data to better predict changes in operations, sales, and customer service. It works both ways.