T-Cube is especially useful for rapidly retrieving responses to ad-hoc queries against large datasets of additive time series labeled using a set of categorical attributes. It can be used as a general tool to support any task requiring access to such data. From the application’s perspective it is transparent: it acts just like the database itself, but an incredibly quickly responding one. The authors had a chance to put T-Cubes into practical use as an enabling technology in applications requiring massive screening of multidimensional temporal data. These applications include two systems to support monitoring of food and agriculture safety and predictive analytics developed at the US Department of Agriculture and the Food and Drug Administration, as well as a system to monitor and forecast health of a fleet of aircraft operated by the US Air Force.
Objective
T-Cube, a data structure designed to efficiently represent large collections of temporal data has been shown to benefit surveillance applications involving monitoring sales of over-the-counter medications and emergency department visits. In this paper we present efficiencies which can be realized in practical applications of T-Cube beyond its original areas of deployment, and we advocate a widespread use of it as a technology which makes manual ad-hoc lookups as well as many kinds of complex automated analyses feasible.