This paper presents a comprehensive data analysis and visualization of electricity consumers’ prepaid bills of Tulkarm district. 250,000 electricity consumers prepaid bills covering the timeperiod from June to December 2018 are analyzed. The application of data mining techniques for understanding electricity consumers’ behavior in electricity consumption and their behavior in charging their electricity meter’s smart cards in terms of quantities charged and charging frequencies in different time periods, areas and tariffs are used. Understanding consumers’ behavior will support planning and decision making at strategic, tactical and operational levels. This analysis is useful for predicting and forecasting future demand with a certain degree of accuracy. Monthly, weekly, daily and hourly time periods are covered in the analysis. Outliers detection using visualization tools such as box plot is applied. K-Means unsupervised machine learning clustering algorithm is implemented. Support Vector Machine classification method is applied. As a result of this study, electricity consumers’ behavior in different areas, tariffs and timing periods is understood and presented by numbers and graphs and new electricity consumer segmentation is proposed

Title

Energies

Publisher

MDPI

Publisher Country

Switzerland

Indexing

Scopus

Impact Factor

2.707

Publication Type

Online only

Volume

12

Year

2019

Pages

22-4287