Datasets¶
A dataset is defined by its Dataset Config. This data structure describes the dataset’s attributes, dimensionality, and file format characteristics.
Types¶
BenchmarkHistoricalModeled
Dimensions¶
Similar to projects, dataset configs can contain already-registered dimension IDs or unregistered dimension configs. Refer to Dimensions and Mappings for background. Dataset contributors will likely want to define new dimensions in the dataset config using the streamlined workflow.
Trivial Dimensions¶
Not all dataset dimensions have to be significant, e.g., historical data will generally have a
trivial (i.e., one-element) scenario dimension. We call these one-element dimensions trivial
dimensions. For space saving reasons, these trivial dimensions do not need to exist in the
dataset files, however, they must be declared as trivial dimensions in the dataset config.
File Format¶
A dataset must comply with a supported dsgrid file format. Please refer to Dataset Formats for options.
Submit-to-Project¶
A dataset must be submitted to a project before it can be used in dsgrid queries. The dsgrid submission process verifies that the dataset’s dimensions either meet the project requirements or have valid mappings.
Examples¶
The dsgrid-StandardScenarios repository contains datasets that you can use as examples.
Historical¶
EIA 861 Utility Customer Sales (MWh) by State by Sector by Year for 2010-2020
Modeled¶
AEO 2021 Reference Case Residential Energy End Use Annual Growth Factors