© 2020, The Joint Center on Global Change and Earth System Science of the University of Maryland and Beijing Normal University.This study was designed to analyse the interaction of household surveys and electricity consumption in Turkey. We used the micro-data set of the Household Budget Survey published by TurkStat for the period 2002–2017. A statistical method and an optimisation method, namely Principal Component Analysis (PCA) and Analytical Hierarchical Processing (AHP), were used to explore the correlations between the components of the household surveys. The first 35 variables among 50 variables were extracted using PCA, and expert opinions validated 26 of 35 ranked using the AHP method. The Artificial Neural Networks (ANN) model, constructed using the input variables defined by expert opinions, gave better prediction results than the ANN model defined using the PCA outcome. ANN sensitivity analysis was conducted to examine the prediction for the components that were validated by AHP evaluations. The results show that dwelling characteristics had more impact on electricity utilisation than did ownership of appliances. It was also discovered that the amount of total expenditure had a negligible impact on electricity consumption.