This data set includes 101 daily summed time series that are labeled for abrupt data shifts. The data, which includes individual sensor-based power and irradiance streams, was collected via the PV Fleets Initiative. It has been normalized and any identifying information has been removed. Raw data is available via the CSV files, and the visualized raw data is plotted in the corresponding HTML files. This data set can be used to train machine learning models to detect abrupt time shifts in time series, as well as for model benchmarking.

Dataset Metadata

Author kperry
Maintainer Email
DOI 10.21948/1874783
Institution NREL
Data Source Type Modeling and Simulation
Method Machine learning
Other Details Labeled sensor-based time series
Owner kperry
Funding SETP 10311-35.01.01

Author kperry
Updated July 7, 2022, 17:38 (UTC)
Created June 8, 2022, 16:33 (UTC)