US Patent 9235630 Dataset discovery in data analytics

An initial work package is obtained. The initial work package defines at least one hypothesis associated with a given data problem, and is generated in accordance with one or more phases of an automated data analytics lifecycle. A plurality of datasets is identified. One or more datasets in the plurality of datasets that are relevant to the at least one hypothesis are discovered. The at least one hypothesis is tested using at least a portion of the one or more discovered datasets.