Data Analytics Help Improve Smart Home Devices

Anyone who has had a smart fire alarm go off for no reason knows smart home products don’t always live up to their billing. Data analytics are helping to reduce such product failures.

Rapid Advances for Data Analytics

Data analytics capabilities have advanced rapidly in recent years. One significant improvement is the use of edge computing, which enables analytics to be collected at the device level without going to servers in a cloud. Edge capabilities are being paired with algorithmic extraction and aggregation of appliance signatures to improve interconnection with utility and weather data streams and provide more rapid and granular data collection. Together, these functions go well beyond diagnostics and reporting—they enable personalized device operation, dynamic use profiling, and reliable safety management for smart homes.

Some products are packaging these capabilities to effectively address threats. For example, preventing electrical fires can be difficult because faulty wiring can remain undetected for long periods. Current fire alarms work only when a fire has already occurred, and preventive devices such as arc-fault circuit interrupters discern problems only on a local circuit. By contrast, Whisker Labs’ Ting smart plug samples electricity and power quality 27 million times per second across all home circuits. It uses machine learning algorithms to aggregate fragmented circuit data from across the home and distinguish dangerous line spark precursors from harmless anomalies. By identifying these precursors, Ting avoids a fire before it happens. The Ting plug sensor also incorporates weather and utility data to form a more complete picture of internal and external factors influencing a home’s power supply. Users manage notifications, live voltage, surge, brownout, and other information via a mobile app.

Regarding energy, data analytics are enabling new efficiencies and improved customer engagement. Typical smart meter data collection occurs at monthly to 15-minute intervals depending on the device. The latest smart meters can provide more granular data, sometimes at per-second intervals. Data analytics provider Bidgely disaggregates 15- and 60-minute smart meter data to the appliance level to construct personalized usage profiles for each home using machine learning algorithms. Bidgely also pairs this usage information with utility data to construct custom load shifting recommendations that support behavioral or incentives programs.

Homeowners and Insurers Take Notice

As devices improve, both homeowners and insurers are taking notice. Nearly 8 in 10 (78%) residential respondents in a recent study by LexisNexis Risk Solutions said they would be willing to share data from smart home devices with their insurance carrier if offered a discount or some incentive.

Building codes do not yet mandate smart home safety devices, but this is likely to change as the benefits of advanced intelligent devices become more widely apparent. For more information about smart home data analytics, please see the forthcoming report from Guidehouse Insights, a Guidehouse company, Market Data: Smart Home Data Analytics.