REDD (Reference Energy Disaggregation Data) dataset

Information

• The REDD (Reference Energy Disaggregation Data) dataset is a publicly available dataset of high-frequency power measurements from multiple homes, intended for use in research on energy disaggregation and related topics.
• The dataset was created in 2011 by the NILM (non-intrusive load monitoring) community, and is hosted by the Electric Power Research Institute (EPRI).
• The REDD dataset contains data from six homes, with measurements taken at sub-minute intervals over a period of several weeks. The homes are located in the United States and are occupied by families with typical energy usage patterns.
• Each home has a different set of appliances and devices, with varying power signatures, making it a useful tool for testing energy disaggregation algorithms.
• The data is provided in CSV format and includes measurements for the total household power consumption as well as individual appliance-level power consumption, where available.
• The dataset also includes metadata describing the homes and appliances, and a detailed documentation on how the data was collected and processed.
• The REDD dataset has been used in numerous research studies and competitions related to energy disaggregation and load profiling, and has contributed to the development of various machine learning and signal processing algorithms.
• The dataset is available for free download and use, but users are required to register and agree to a data use policy that includes restrictions on commercial use and redistribution of the data.