Efficient production and consumption of electricity have become the focal point of modern civilization and industrialization. Due to economic and political reasons, we have yet to find any better yet scalable technology to find the next best alternative to the massively inefficient way of producing electricity with fossil fuel. In this task, I propose alleviating this problem by focusing on electricity consumption and exploring ways to find what modern design can offer. By exploring Site Energy Usage Intensity (EUI) on various stages and scales, I have identified sectors that contribute most to the inefficiencies. I also researched the various extraneous relevant factors that have some effects on it. Parameterizing these variables, I have modeled the prediction of site EUI using basic machine learning models such as Linear Regression and Gradient Boosting. Tuning these models' hyperparameters using standardized methods yields a highly reliable prediction of EUI. Finally, I have experimented with feature importance analysis to test the validity of our hypothesis. This experiment elicits further improvement scope of the proposed system, which is crucial to identify the most important factors to focus our research on the efficiency of electricity consumption.