The Internet of Things (IoT) has been a buzzword for a few years now. Companies are moving to production with various use cases, but most of them are around machines and logistics. A much less popular IoT use case is around understanding human behavior through biometrics. Proper use of biometrics enables companies to fine-tune the most expensive and most impactful part of the organization — the people.
Business Use Cases
Let’s make this tangible. If you’re managing a team of skilled professionals in charge of running critical machinery—say, airline pilots or construction workers—being able to detect someone’s apparent drowsiness or other attention deficit can make the difference between a crash and a safe day. In some of the applications I’ve seen, biometric data has been able to predict when someone is about to fall ill and should not be behind the wheel so to to avoid a possible major accident.
There are multiple biometric markers that can be obtained for this. Heart rate variability speaks volumes to the emotional state. Add information on body temperature, pupil dilation, blood oxygenation, and related physiological data — all easily measurable today — and we have quite a bit of data on what’s happening internally. Along with predictions based on individual patterns, we can tell someone to take the last 3 hours of the day off if we see that his or her accident forecast is going up with the pending flu.
Another dimension is actually predicting health issues. We’ve all heard of how the Apple Watch can help predict heart attacks. In the same vein, the Oracle Santa Monica Cloud Solution Hub just built a technical concept demo on Atrial fibrillation (A-Fib) detection and how to best start the treatment based on basic data from commercially available wearable sensors*. While turning these into real clinical tools is still a ways away, being able to monitor the population at risk instead of a few individuals who are known to have the diagnosis can make a massive difference for the future of preventative medicine.
Being able to collect the data is one thing, but being able to analyze the data and turn the blips on the screen into meaningful information, insights, and, finally, actions is even more critical. Using a robust stream analytics platform like Oracle Stream Analytics, Oracle Analytics Cloud, and Oracle Artificial Intelligence is a great way to jump this process and start discovering trends. For example, streaming all biometric data to the Stream Analytics will allow you to clean and analyze it on the fly, add alerts for known issues, build trends and predictions, and store it in a data lake for future analysis and pattern discovery.
While measuring the biometrics and doing predictions can feel a bit out there, it is all very much doable today. That being said, there is another major component to pay attention to: digital ethics. While measuring these parameters, are we learning something that we might not expect? Detecting that someone is about to have a heart attack is good information to have and act on, but detecting that someone just got pregnant may not be data that we want to have if we want to avoid privacy and discrimination issues. It’s critical to build policies and safeguards on what can be learned from the data and how it is used in order to gain the trust of the user population and population at large. Making sure your data retention and anonymization policies work in such a way that the data is mostly meaningless even if a leak occurs is paramount in doing this right.
*Full disclosure: no clinical trials have been carried out as we're simply testing out the ability to use the data for large-scale diagnosis.