In this empirical study, socioeconomic factors that can easily be extracted from families have been used to build a "home electricity usage prediction" model based on two variables, family monthly income and family size. Each of these factors was evaluated individually. Two machine learning models were built using those factors as features. Two new parameters were introduced to analyze the data. Models are based on "Linear Regression" and "Random Forest" algorithms. This study revealed that the socioeconomic factors such as family size and family income are very effective in domestic electricity usage prediction model building, where the end usages are not known. Furthermore, the random forest algorithm was found to be more effective for unseen data than the linear regression algorithm. The accuracy of the models can be further improved by adding more data into the both models.