Society’s energy consumption has shot up in recent years, making the prediction of its demand a current challenge to ensure an efficient and responsible use. Artificial intelligence techniques have proven to be potential tools in handling tedious tasks and making sense of large-scale data to make better business decisions in different areas of knowledge. In this article, the use of random forests algorithms in a Big Data environment is proposed for households energy demand forecasting. The predictions are based on the use of information from different sources, confirming a fundamental role of socioeconomic data in consumer’s behaviours. On the other hand, the use of Big Data architectures is proposed to perform horizontal and vertical scaling of the solution to be used in real environments. Finally, a tool for high-resolution predictions with great efficiency is introduced, which enables energy management in a very accurate way. Raw data is incuded in data.csv. This file contains half hourly home electricity consumption registers for 4404 households with fix tariffs (not subject to dynamic time of use) for a period between November 2011 and February 2014. Original information was acquired from the Low Carbon London project led by UK Power Networks (https://data.london.gov.uk/dataset/smartmeter-energy-use-data-in-london-households) RFResults.zip contains the energy predictions for each ACORN group using the generated Random Forest algorithm. For this purpose, the first 613 days of a total of 818 observations of each group were considered for training and the last 205 days for testing. Meteorological data was adquired from the darksky app (https://darksky.net). These data are included in the weather_hourly_darksky.csv uk_bank_holidays. xlsx contains the dated of UK bank holidays for the studied period, used as additional variable related to occupancy

Abstract: Measurements of electric power consumption in one household with a one-minute sampling rate over a period of almost 4 years. Different electrical quantities and some sub-metering values are available.

Source: Georges Hebrail (georges.hebrail '@' edf.fr), Senior Researcher, EDF R&D, Clamart, France Alice Berard, TELECOM ParisTech Master of Engineering Internship at EDF R&D, Clamart, France

Data Set Information: This archive contains 2075259 measurements gathered in a house located in Sceaux (7km of Paris, France) between December 2006 and November 2010 (47 months). Notes:

(global_active_power*1000/60 - sub_metering_1 - sub_metering_2 - sub_metering_3) represents the active energy consumed every minute (in watt hour) in the household by electrical equipment not measured in sub-meterings 1, 2 and 3. The dataset contains some missing values in the measurements (nearly 1,25% of the rows). All calendar timestamps are present in the dataset but for some timestamps, the measurement values are missing: a missing value is represented by the absence of value between two consecutive semi-colon attribute separators. For instance, the dataset shows missing values on April 28, 2007.

Attribute Information:

date: Date in format dd/mm/yyyy time: time in format hh:mm:ss global_active_power: household global minute-averaged active power (in kilowatt) global_reactive_power: household global minute-averaged reactive power (in kilowatt) voltage: minute-averaged voltage (in volt) global_intensity: household global minute-averaged current intensity (in ampere) sub_metering_1: energy sub-metering No. 1 (in watt-hour of active energy). It corresponds to the kitchen, containing mainly a dishwasher, an oven and a microwave (hot plates are not electric but gas powered). sub_metering_2: energy sub-metering No. 2 (in watt-hour of active energy). It corresponds to the laundry room, containing a washing-machine, a tumble-drier, a refrigerator and a light. sub_metering_3: energy sub-metering No. 3 (in watt-hour of active energy). It corresponds to an electric water-heater and an air-conditioner.

Relevant Papers: N/A

Citation Request: This dataset is made available under the “Creative Commons Attribution 4.0 International (CC BY 4.0)” license

based on the ESP32 hardware.

To allow for a proper assessment of the reliability contribution of renewables on any grid, it is imperative that the electricity demand is reflective of actual consumption. Historical national demand time series profiles are not publicly available from Nigeria’s transmission system operator; therefore, a long-term demand profile had to be developed. Using hourly demand timeseries of 5 Nigerian utilities, with an aggregate regulated offtake requirement of 53.5% of total grid generation, we have reconstructed a timeseries demand "suppressed" for the entire Nigerian grid. This analysis was performed for a five-year period and validated against annual energy consumption data from the multi-year tariff order (MYTO) financial model used by the Nigerian Electricity Regulatory Commission (NERC). Validation results yield an average error margin of 5.8%, while the peak demand for period is 4.8GW. To establish the hourly national unsuppressed demand profile, the unsuppressed demand profile for each utility is first generated, and then the national profile is estimated based on an historical network-wide stress test. Due to significant load shedding events, the measured timeseries demand data of utilities typically are populated with blanks. Therefore, a suitable method is required to estimate the demand that would have otherwise occurred if there were no outage events. A structural model developed to reconstruct unsuppressed demand data based on the historical observations of suppressed demand on the feeder is applied to the utility demand data. The reconstructed demand timeseries leverages data either side of a load shedding event to fill in a likely value resulting in better estimates of demand at times of the day when a feeder is often disconnected. It is then possible to aggregate the reconstructed timeseries profiles to determine the unsuppressed grid connected demand. The peak unsuppressed demand for the period is 8.0GW, with an average unmet energy demand of 51.5% (26.7TWh). The 2016 timeseries data provided here is for research purposes only.

According to a 2024 forecast, global electricity consumption of data centers was projected to grow from 330 terawatt-hours in 2022 to over one petawatt-hour in 2030. This would represent around 3.7 percent of the total electricity consumption worldwide by the end of the period under consideration. Artificial intelligence accounted for around 4.5 percent of the data centers' electricity consumption in 2023. This figure is projected to grow over the next five years.

This Dataset represents the best consumption information up to date, presenting a forecast until the end of the year.The Dataset is composed of the following elements:Energy consumption of installations with MAT voltage level (very high voltage);Energy consumption of installations with AT voltage level (high voltage);Energy consumption of installations with MT (medium voltage) voltage level;Energy consumption of installations with BT voltage level (low voltage);Sum of energy consumption of all voltage levels (Total).The data is obtained through the best information on existing consumption until the data, making a forecast until the end of the year based on historical values. At the end of each 15-minute period, the total amount corresponds to the energy during that interval.The update frequency is monthly.Use case: This data allows you to monitor the best information on national energy consumption, quarter-hourly by voltage level during the year in question.1. The information made available by E-REDES constitutes an approximation to the values taken from the system and is based on the moment in which it is collected. Given that the connection points, the electricity distribution network, and the consumption and production values themselves are naturally very dynamic, it is safeguarded that the information made available may be subject to subsequent changes and updates, with the exception of any omissions and/or occasional inaccuracies of location that the information may contain.2. In this way, E-REDES is not liable to third parties, namely, partners, service providers, contractors, users and customers, for damages that may arise as a result, direct or indirect, of the use of this Information, in particular when carrying out interventions, calculations and/or estimates, without confirming the accuracy and updating of the data, whereby it is duly noted that the consultation of this information does not affect the duty to promote a direct consultation with E-REDES in order to obtain updated information.The data provided by the E-REDES Open Data Portal is covered by open licenses (CC BY 4.0). There are no restrictions on access, under the commitment that data users cite the publisher. Therefore, we suggest that you cite the Open Data E-REDES Portal as:E-REDES – Distribuição de Eletricidade, “E-REDES Open Data Portal”. Accessed in “Data”. [Online] Available at https://e-redes.opendatasoft.com/pages/homepage/If you share on social media, please add #PortalOpenData_E_REDES

The electricity consumption model outputs an estimate of electricity consumption at the Swiss national scale. The model is trained on data from the following sources: hourly national consumption provided by Swissgrid (Energy overview), meteorological data from MeteoSwiss (including weather forecast), and calendar information such as public holidays as provided by the python holidays package and manually acquired holidays. The model belongs to the category of Generalized Additive Models (GAMs) and output is a prediction at an hourly and daily basis with a lower and upper confidence interval of 2.5 and 97.5 respectively.

This dataset is the first release of data for the 2025 Competition on Electric Energy Consumption Forecast Adopting Multi-criteria Performance Metrics

The competition is open and welcomes everyone who wishes to participate and to anyone who can benefit from these data.

Competition Outline

Forecasting of electric energy consumption can be a very difficult tasks when handling building-level data. However, an accurate forecast is needed to boost the potential of energy management systems. The need to forecast energy consumption grows as our reliance on renewable energy sources, such as solar and wind power, grows. This means that to meet consumer demand with renewable energy generation, energy management systems must operate based on accurate energy forecasting models for both short and long-term periods. Energy consumption forecasting techniques that can manage a variety of scenarios, including varying prediction timeframes, accessible data, data frequency, and even data quality, have been the subject of intense research. There is no one-size-fits-all approach, where certain situations call for different approaches. The goal of this competition is to compile and evaluate the most recent advances in energy consumption forecasting techniques.

Releases Details

• v1.0: one year of data from a smart building with readings taken every 5 minutes.
• v2.0: 40 days of data from a smart building with readings taken every 5 minutes.
• v3.x: a single day of data from a smart building with readings taken every hour. These releases will become available during the first competition period (from 06/01/2025 to 10/01/2025).
• v4.x (to be confirmed): a single day of data from a smart building with readings taken every hour. These releases will become available during the second competition period.

Dataset Description

All releases are composed of the following data:

• Time: in hours and minutes
• Power: in Watts
• Voltage: in Volts
• Current: in Ampers
• Generation power: in Watts
• Temperature: in ºC

Demand for electric power, especially amidst limited fossil fuel-based generation capacity, has elevated renewable energy sources to a forefront solution for the growing energy needs. Solar energy, a key renewable source through photovoltaic (PV) panels, faces challenges such as intermittency and non-dispatchability. Thus, recent research has focused on developing programs to predict near-future solar energy generation, with machine learning being a pivotal approach. This article details the creation of an effective machine-learning pipeline for predicting future hourly power generation based on weather data (e.g. temperature, humidity, irradiance). The pipeline, aimed at a scheduling system in a farm equipped with a Solar Power System (SPS) in Al-Salt, Jordan, was optimized using Genetic Algorithm and Grid Search methods. The objective of this article is to create an optimal pipeline with minimal loss. The evaluation shows that ensemble regressors, especially Gradient Boosting Regressors, are effective. This is evidenced in the grid search pipeline, which outperformed the TPOT optimization pipeline-derived pipeline, the latter including stacked ensemble regressors and sequential preprocessors.

According to a recent forecast, global electricity consumption will range between ****** and ****** terawatt-hours in 2030 and between ****** and ****** terawatt-hours in 2050, depending on the energy transition scenario. However, to keep the global temperature increase below *** degrees Celsius, the total electricity consumption in the world should be at ****** terawatt-hours in 2030, a result that is not likely to be achieved even if the current energy transition commitments are achieved (Achieved Commitments scenario).

Data and Problem Description

More accurate forecasts of building energy consumption mean better planning and more efficient energy use. So The objective is to forecast energy consumption from following data: (For each data set, several test periods over which a forecast is required will be specified.)

Historical Consumption

A selected time series of consumption data for over 260 buildings.

-obs_id - An arbitrary ID for the observationaa -SiteId - An arbitrary ID number for the building, matches across datasets -ForecastId - An ID for a timeseries that is part of a forecast (can be matched with the submission file) -Timestamp - The time of the measurement -Value - A measure of consumption for that building

Building Metadata

Additional information about the included buildings.

-SiteId - An arbitrary ID number for the building, matches across datasets -Surface - The surface area of the building -Sampling - The number of minutes between each observation for this site. The timestep size for each ForecastId can be found in the separate "Submission Forecast Period" file on the data download page. -BaseTemperature - The base temperature for the building -IsDayOff - True if DAY_OF_WEEK is not a work day

Historical Weather Data

This dataset contains temperature data from several stations near each site. For each site several temperature measurements were retrieved from stations in a radius of 30 km if available. Note: Not all sites will have available weather data.

Note: Weather data is available for test periods under the assumption that reasonably accurate forecasts will be available to algorithms that the time that we are attempting to make predictions about the future.

-SiteId - An arbitrary ID number for the building, matches across datasets -Timestamp - The time of the measurement -Temperature - The temperature as measured at the weather station -Distance - The distance in km from the weather station to the building in km

Public Holidays

Public holidays at the sites included in the dataset, which may be helpful for identifying days where consumption may be lower than expected.Note: Not all sites will have available public holiday data.

-SiteId - An arbitrary ID number for the building, matches across datasets -Date - The date of the holiday -Holiday - The name of the holiday

Acknowledgements

Forecasting energy consumption data published by Schneider Electric.

Inspiration

Three time horizons and time steps are distinguished for more than 260 building sites are provided. The goal is either:

• To forecast the consumption for each quarter for the length of time specified by the submission format.
• To forecast the consumption for each hour for the length of time specified by the submission format.
• To forecast the consumption for each day for the length of time specified by the submission format.

Historical data are given at the granularity that is required for the consumption forecast. So, when historical data are given by steps of 15 minutes, forecasts are required by steps of 15 minutes. When historical data are given by steps of 1 hour, forecasts are required by steps of 1 hour. When historical data are given by steps of 1 day, forecasts are required by steps of 1 day.

#

*If you find the data useful your upvote is an explicit feedback for future works, Have fun exploring data!*

Context

Experimental data used to create regression models of appliances energy use in a low energy building.

Content

Data Set Characteristics:

Multivariate, Time-Series, Regression

Number of Instances(Rows):

19735

Number of Attributes(Columns):

29

Associated Tasks:

Regression

Source:

Luis Candanedo, luismiguel.candanedoibarra '@' umons.ac.be, University of Mons (UMONS).

Data Set Information: Given in Metadata tab about the sources and collection methodology.

Attribute Information:

date time year-month-day hour:minute:second

Appliances, energy use in Wh (target variable for prediction)

lights, energy use of light fixtures in the house in Wh

T1, Temperature in kitchen area, in Celsius

RH_1, Humidity in kitchen area, in %

T2, Temperature in living room area, in Celsius

RH_2, Humidity in living room area, in %

T3, Temperature in laundry room area

RH_3, Humidity in laundry room area, in %

T4, Temperature in office room, in Celsius

RH_4, Humidity in office room, in %

T5, Temperature in bathroom, in Celsius

RH_5, Humidity in bathroom, in %

T6, Temperature outside the building (north side), in Celsius

RH_6, Humidity outside the building (north side), in %

T7, Temperature in ironing room , in Celsius

RH_7, Humidity in ironing room, in %

T8, Temperature in teenager room 2, in Celsius

RH_8, Humidity in teenager room 2, in %

T9, Temperature in parents room, in Celsius

RH_9, Humidity in parents room, in %

To, Temperature outside (from Chievres weather station), in Celsius

Pressure (from Chievres weather station), in mm Hg

RH_out, Humidity outside (from Chievres weather station), in %

Wind speed (from Chievres weather station), in m/s

Visibility (from Chievres weather station), in km

Tdewpoint (from Chievres weather station), °C

rv1, Random variable 1, nondimensional

rv2, Random variable 2, nondimensional

Where indicated, hourly data (then interpolated) from the nearest airport weather station (Chievres Airport, Belgium) was downloaded from a public data set from Reliable Prognosis, rp5.ru. Permission was obtained from Reliable Prognosis for the distribution of the 4.5 months of weather data.

Acknowledgements

Luis M. Candanedo, Veronique Feldheim, Dominique Deramaix, Data driven prediction models of energy use of appliances in a low-energy house, Energy and Buildings, Volume 140, 1 April 2017, Pages 81-97, ISSN 0378-7788, Web Link.

Citation

Dua, D. and Graff, C. (2019). UCI Machine Learning Repository [http://archive.ics.uci.edu/ml]. Irvine, CA: University of California, School of Information and Computer Science.

Inspiration

1) This is a regression task, You should predict the "appliances" column. Column descriptions are given above. Please read them before proceeding. 2) Appropriate time series analysis with regression is preferred more. 3) Exploratory data analysis with charts and plots.

Have fun!

Electricity consumption prediction plays a vital role in intelligent energy management systems, and it is essential for electricity power supply companies to have accurate short and long-term energy predictions. In this study, a deep-ensembled neural network was used to anticipate hourly power utilization, providing a clear and effective approach for predicting power consumption. The dataset comprises of 13 files, each representing a different region, and ranges from 2004 to 2018, with two columns for the date, time, year and energy expenditure. The data was normalized using minmax scalar, and a deep ensembled (long short-term memory and recurrent neural network) model was used for energy consumption prediction. This proposed model effectively trains long-term dependencies in sequence order and has been assessed using several statistical metrics, including root mean squared error (RMSE), relative root mean squared error (rRMSE), mean absolute bias error (MABE), coefficient of determination (R2), mean bias error (MBE), and mean absolute percentage error (MAPE). Results show that the proposed model performs exceptionally well compared to existing models, indicating its effectiveness in accurately predicting energy consumption.

The energy demand from the United States' electric vehicle fleet amounted to **** terrawatt per hour in 2021. Demand is forecast to grow to *** terrawatt-hours in 2030, when the U.S. electric vehicle parc will account for **** million vehicles. Passenger cars are the segment requiring the most energy.

Predictions of energy consumption are crucial for energy retailers to minimize deviations from energy acquired in the day-ahead market and the actual consumption of their customers. The increasing spread of smartmeters means that retailers have access to hourly consumption values of all their contracted customers in realtime. Using machine learning algorithms, these hourly values can be used to calculate predictions for the future energy consumption of the customers. The present data set allows the training and validation of AI-based prediction models.

In 2022, electricity demands of over 22.7 thousand gigawatt hours were expected in Ghana, increasing from an estimated amount of 21.3 thousand gigawatt hours in the previous year. By 2030, it was forecast to rise even further, reaching approximately 36.5 thousand gigawatt hours.

This is a dataset used in EnergyDetective 2020, which is a competition on building energy consumption prediction hold in 2020 by Xupeng Research Group.

With the prediction case provided in the competition, which is to predict the energy consumption of a building with some physical information but without historical data of its own, the dataset will include two parts, which is the test building dataset and the reference building dataset. The test building dataset includes more data about the physical description for the building while the reference building dataset includes more historical consumption data.

All the hourly energy consumption data comes from office buildings located in Shanghai, China during 2015-2017. Energy consumption data is divided into two meter type in the data set. Energy cost by lights and plugs is one of the meter type (marked as “Q”) and that cost by the HVAC system is the other(marked as “W”). The consumption records of two meter types are gathered by raw meta data, which comes from a private dataset built by an energy management company.

In the meanwhile, weather data is given in the dataset. And the weather data is collected by a real weather station in Shanghai, China.

In the new version, we provide the historical hourly energy consumption data of the test building.

Nigeria Electricity Consumption: Average per Hour data was reported at 4,105.660 MWh in Dec 2024. This records an increase from the previous number of 4,000.240 MWh for Sep 2024. Nigeria Electricity Consumption: Average per Hour data is updated quarterly, averaging 2,976.960 MWh from Mar 2005 (Median) to Dec 2024, with 79 observations. The data reached an all-time high of 4,456.000 MWh in Sep 2015 and a record low of 1,300.000 MWh in Jun 2008. Nigeria Electricity Consumption: Average per Hour data remains active status in CEIC and is reported by Central Bank of Nigeria. The data is categorized under Global Database’s Nigeria – Table NG.RB002: Electricity Generation and Consumption.

Electricity consumption prediction plays a vital role in intelligent energy management systems, and it is essential for electricity power supply companies to have accurate short and long-term energy predictions. In this study, a deep-ensembled neural network was used to anticipate hourly power utilization, providing a clear and effective approach for predicting power consumption. The dataset comprises of 13 files, each representing a different region, and ranges from 2004 to 2018, with two columns for the date, time, year and energy expenditure. The data was normalized using minmax scalar, and a deep ensembled (long short-term memory and recurrent neural network) model was used for energy consumption prediction. This proposed model effectively trains long-term dependencies in sequence order and has been assessed using several statistical metrics, including root mean squared error (RMSE), relative root mean squared error (rRMSE), mean absolute bias error (MABE), coefficient of determination (R2), mean bias error (MBE), and mean absolute percentage error (MAPE). Results show that the proposed model performs exceptionally well compared to existing models, indicating its effectiveness in accurately predicting energy consumption.

The total electric power consumption in Indonesia was forecast to continuously increase between 2024 and 2029 by in total **** million kilowatt hours (+10 percent). The electric power consumption is estimated to amount to **** million kilowatt hours in 2029. Depicted is the estimated electric power consumption per capita in the country or region at hand. Both demand from private households as industrial consumption are included in the figures.The shown data are an excerpt of Statista's Key Market Indicators (KMI). The KMI are a collection of primary and secondary indicators on the macro-economic, demographic and technological environment in up to *** countries and regions worldwide. All indicators are sourced from international and national statistical offices, trade associations and the trade press and they are processed to generate comparable data sets (see supplementary notes under details for more information).Find more key insights for the total electric power consumption in countries like Cambodia and Myanmar.