Data Sets

A data set is collection of signals, such as price and revenue, those signals’ time series and a graph of entities, starting from entities having those time series.

A data set is uniquely defined by the collection of its signals. The graph of entities in the data set is built following relationships of ownership/aggregation type, for example HAS_BRAND, coming from other entities into entities with corresponding time series, and expanded further with ownership relations coming into the resulting graph, etc until no more ownership relations can be found. (An ownership relationship type has its isOwnership field set to true.)

The graph of a data set is not necessarily connected.

There are no data sets in the global namespace. Customers can only create data sets in their own namespace.

All signals of a data set must exist when a data set is created. The collection of signals in a data set can be modified later. The signals are to be treated as a set, i.e. signals are not repeated and can be inserted or returned in any order.