![]() ![]() This repository with new cleaned data annually.ĭata is stored in csv format with each row corresponding to an hour of demand information. Is a continuously growing data record in the EIA database, we plan to update The final data product is available to everyone. The original data from the 10 September 2019 data query is located in directory data/release_2019_Oct/original_eia_files/Īnd can be used to compare results or analyze other methods of cleaning. The reported demand value for each hour corresponds to the integrated mean value in megawatts over the previous hour. The EIA began collecting hourly demand data in July of 2015 and continuously publishes new values each day. The 11 balancing authorities that include generation only are denoted as such. The specific hourly demand data used is originally located here:Īt the end of this README is a list of the 67 balancing authorities in the contiguous U.S.ĭemand data is provided for the 56 balancing authorities which have demand. Raw data is made available to the public through the EIA Open Data portal documented here: Energy Information Administration (EIA) via Form-930. Original hourly demand data is collected from electric balancing authoritiesīy the U.S. The data span exactly 4 years, from the start of 2 July 2015 through the end of 1 July 2019. The first day of data has been removed because of significant reporting inconsistencies. Spans from their initial data entries on 05:00:00 UTC to 23:00:00 UTC. The raw data were queried from the EIA database on Septemand EIA Cleaned Hourly Electricity Demand Data (Version v1.1). Developing reliable hourly electricity demand data through screening and imputation. The MICE technique provides complete data sets without any extremely anomalous values.įull documentation of the cleaning process has been published in Scientific Data. Multiple Imputation by Chained Equations (MICE) technique. We impute missing demand values and the values flagged as anomalous using a Which constitute about 0.5% of the total data. We have developed a data cleaning process that consists of flagging anomalous demand values, Energy Information Administration (EIA) show 2.2% of hourly values are missing based The raw hourly electricity demand data queried from the Subregion level data for the balancing authorities that provide this and provideĢ full years of cleaned data spanning 1 October 2018 through 30 September 2020 in theĭata/release_2020_Oct_include_subregions/ directory. In data/release_2020_Oct/ and corresponds to 5 full years of cleaned data spanningįrom 2 July 2015 through 1 July 2020 and follow the methods discussed in the paper exactly.Īdditionally, we have modified the method to incorporate The most current cleaned data can be found The cleaned data have been updated since publication of the methods paper (early 2020ĭiscussed throughout this repository. EIA_Cleaned_Hourly_Electricity_Demand_DataĪ repository for publishing and versionsing cleaned EIA hourly demand dataįind archived versions of the code used to clean this data: Most current data releases
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |