WIT Press


GEOGRAPHICALLY WEIGHTED PRINCIPAL COMPONENTS ANALYSIS APPROACH TO EVALUATE ELECTRICITY CONSUMPTION BEHAVIOUR

Price

Free (open access)

Volume

249

Pages

12

Page Range

101 - 112

Published

2020

Paper DOI

10.2495/SC200091

Copyright

Author(s)

NATHALIA TEJEDOR-FLORES, PURIFICACIÓN VICENTE-GALINDO, PURIFICACIÓN GALINDO-VILLARDÓN

Abstract

Electricity consumer behaviour is primarily based on individual decisions, which are often driven by external factors such as economic incentives, existing demographics, environmental variables, social norms, and infrastructure. This study aims to understand the amount of electricity used per hour dedicated in the household (HH) sectors and the paid sectors – including agriculture (AG), other productive sectors (PS) and the service and government sector (SG) – using Geographically Weighted Principal Components Analysis (GWPCA). In the literature, we found that a standard Principal Components Analysis (PCA) can be replaced with a GWPCA when we want to account for a certain spatial heterogeneity. To use the GWPCA to compare the results with a standard PCA, we took the data used in a previous investigation and applied both analyses in order to find a better way to understand the electricity consumer behaviour, using a multivariate analysis. The standard PCA reveals that the first three components collectively account for 73.66% of the variation in the data. Using GWPCA, we found a clear geographical variation in the percentage of total variance data, with higher percentages (90%–95%) located in the south-west and a small part of the north-east of the case of study used. Also, the electrical Energy Throughput in Paid Work (ETPW), and the amount of energy used per hour dedicated to the Paid Work sector (EMRPW), appears to play an important part in defining the local structure in the south-west (coastal region) and in the northern part of the case of study used, respectively. The comparison results suggest that GWPCA provides better fitness than the standard PCA model by considering spatial heterogeneity.

Keywords

Geographically Weighted Principal Components Analysis, multivariate analysis, sustainable development, electricity consumption