Smart meters have revolutionized the way we monitor energy consumption. These innovative devices have transformed the traditional electricity metering landscape, bringing the efficiency and accessibility of digital technology to homes and businesses. This evolution is not only changing the way we interact with energy, but also deepening our understanding of its use.

More than just recorders of electricity usage, these meters are vital communication links between consumers and utilities. By facilitating real-time data exchange, smart meters provide unparalleled insight into electricity usage patterns. This benefits consumers, who can potentially receive detailed, easy-to-understand reports that encourage more efficient energy use, and utilities, which benefit from improved grid management and billing processes. 

Our team delved into the world of smart meter datasets and uncovered a critical gap in existing research. We found that while there are about a hundred smart meter datasets available to researchers, the majority are from Western, English-speaking countries. This geographic and cultural homogeneity poses a risk of bias, particularly in studies related to human behavior and energy consumption patterns. Addressing this limitation, we have developed a dataset that contributes a unique cultural and geographic perspective, diversifying the current range of smart meter data.

Introducing a pioneering electricity smart meter dataset from Spanish households

Building on the need for a more geographically diverse dataset in energy research, we present our groundbreaking contribution: a comprehensive smart meter dataset from Spain. This dataset is the result of our collaboration in the EU-funded WHY project, which focuses on implementing causal models to analyze everyday energy consumption decisions and responses to interventions.

The cornerstone of our dataset is GoiEner, a forward-looking retail cooperative specializing in renewable energy, which was part of the consortium. Through this partnership, we have gained access to anonymized hourly electricity demand data from an extensive network of 25,559 customers. This diverse group includes not only individual households, but also retail and department stores, offices, industrial facilities, and public facilities. The data covers all provinces in mainland Spain, with a notable concentration in the northern regions of the Basque Country, where GoiEner is based, and Navarre.

What makes this dataset particularly valuable to researchers is its size and scope. It is the first time that an electricity smart meter dataset from Spain has been made publicly available. This is an important step in reducing geographic bias in energy research. The dataset is also one of the largest collections of time series data available. It covers the period from November 2014 to June 2022, providing an average time series duration of approximately three years per customer. This extensive time frame is critical for understanding long-term patterns and trends in electricity consumption.

Insights into consumption patterns

The true potential of our dataset lies in its application to real-world scenarios, particularly in the analysis of household electricity behavior. One of the key aspects of our research is the segmentation of GoiEner's user data based on various criteria, including economic activity. This granularity allows for a nuanced understanding of electricity consumption patterns, especially when focusing on households.

By isolating household data and studying it before the disruptive events of 2020, we gain valuable insights into the routines and habits that characterize Spanish household electricity use. Because Spain experienced one of the most restrictive lockdowns in Europe, this pre-pandemic data is even more important, as it provides a baseline against which to measure the impact of the pandemic on household behavior.

Our analysis shows interesting patterns in residential electricity consumption. On weekdays, there are three distinct peaks: in the morning around 10:00, in the afternoon around 15:00, and a pronounced peak in the late evening around 21:00. These peaks coincide with routine household activities such as cooking, showering, and media use. The weekend data show a contrast, with a noticeable shift in consumption patterns. Activities start later and the distinct weekday peaks become less pronounced. This shift reflects different social practices and routines during the weekend. Interestingly, the analysis is consistent with research by other authors who have analyzed household electricity use and active occupancy profiles using data from the Spanish Time Use Survey.

This analysis is just the tip of the iceberg. The richness of the dataset allows for a variety of behavioral analyses across different economic activities and time periods. By examining these patterns, we can understand not only how much electricity is consumed, but also the underlying behaviors and routines that drive that consumption.

Leveraging data for a sustainable future

The valuable insights from our comprehensive smart meter dataset go far beyond academic interest. They have significant potential to shape future energy policy, urban development, and sustainability efforts. By delving into the detailed patterns of electricity use in homes and businesses, this dataset enables the creation of energy-saving strategies that are more closely aligned with real-world behavior. This is critical to advancing our global efforts to combat climate change and move toward more sustainable energy practices.

In addition, this dataset opens up new opportunities for energy economics and behavioral research. It provides a robust basis for analyzing how various factors, from economic shifts to technological advances to policy changes, affect electricity consumption. The data's potential for predictive modeling is particularly exciting. Using advanced machine learning techniques, it can help predict future energy demand, improve grid management, and facilitate the integration of renewable energy sources.

After all, this dataset is more than just a bunch of numbers: it's a valuable tool for deepening our understanding of energy use and supporting the development of more sustainable and resilient energy infrastructures.