Investigation of the Impact of Economic and Social Factors on Energy Demand through Natural Language Processing

The relationship between energy demand and variables such as economic activity and weather is well established. However, this paper aims to explore the connection between energy demand and other social aspects, which receive little attention. Through the use of natural language processing on a large news corpus, we shed light on this important link. This study was carried out in five regions of the UK and Ireland and considers multiple horizons from 1 to 30 days. It also considers economic variables such as GDP, unemployment and inflation. We found that: 1) News about military conflicts, transportation, the global pandemic, regional economics, and the international energy market are related to electricity demand. 2) Economic indicators are more important in the East Midlands and Northern Ireland, while social indicators are more useful in the West Midlands and the South West of England. 3) The use of these indices improved forecasting performance by up to 9%.

Electricity Demand Forecasting through Natural Language Processing with Long Short-Term Memory Networks

Electricity demand forecasting is a well established research field. Usually this task is performed considering historical loads, weather forecasts, calendar information and known major events. Recently attention has been given on the possible use of new sources of information from textual news in order to improve the performance of these predictions. This paper proposes a Long and Short-Term Memory (LSTM) network incorporating textual news features that successfully predicts the deterministic and probabilistic tasks of the UK national electricity demand. The study finds that public sentiment and word vector representations related to transport and geopolitics have time-continuity effects on electricity demand. The experimental results show that the LSTM with textual features improves by more than 3% compared to the pure LSTM benchmark and by close to 10% over the official benchmark. Furthermore, the proposed model effectively reduces forecasting uncertainty by narrowing the confidence interval and bringing the forecast distribution closer to the truth.

A Quantitative Exploration of Natural Language Processing Applications for Electricity Demand Analysis

The relationship between electricity demand and weather has been established for a long time and is one of the cornerstones in load prediction for operation and planning, along with behavioral and social aspects such as calendars or significant events. This paper explores how and why the social information contained in the news can be used better to understand aggregate population behaviour in terms of energy demand. The work is done through experiments analysing the impact of predicting features extracted from national news on day-ahead electric demand prediction. The results are compared to a benchmark model trained exclusively on the calendar and meteorological information. Experimental results showed that the best-performing model reduced the official standard errors around 4%, 11%, and 10% in terms of RMSE, MAE, and SMAPE. The best-performing methods are: word frequency identified COVID-19-related keywords; topic distribution that identified news on the pandemic and internal politics; global word embeddings that identified news about international conflicts. This study brings a new perspective to traditional electricity demand analysis and confirms the feasibility of improving its predictions with unstructured information contained in texts, with potential consequences in sociology and economics.