The Forest & the Trees: Leveraging OmniSci to Support Geo-Semantic Pattern Detection in Large, Globally Distributed Data Sets

Skyhook Uses Omnisci for Insights in its High Volume, Globally Distributed Mobile Device Datapool Over the past decade and a half Skyhook has built and patented much of the technology used for accurately positioning mobile devices. This is accomplished by mixing Wi-Fi with cell and gps signals to deliver a best hybrid solution with extremely low latency regardless of local sampling challenges. Skyhook pioneered precision location for both iOS and Android and has deployed various solutions across billions of devices in these and other OS pools. In order to remain functionally accurate a global Wi-Fi database must self-heal through active field acquisition, and this requires massive MAU/DAU/HAU activity sustained indefinitely and retained as a structured artifact across multiple years. Skyhook’s various databases, including Wi-Fi access points, cell antennas, ip detections, signal confirmed venues and individual returned device locations produce high temporally and positionally accurate samples tens of billions of times daily. This raw output is refined and fractioned into specific data products that are useful to big data consumers in multiple industries and governments. Omnisci has proven itself highly performant in managing large volumes of point-based, time- and place-anchored data in a low latency analysis environment. We will show examples in areas where GPU-resident analytics permit Skyhook to scale and leverage vast data extracts into actionable insight in areas of crowd characterization, crisis-driven population dynamics, commercial consumer behavior and system pathology detection.