The rapid proliferation of smart meters in residential and commercial buildings has led to the generation of vast amounts of energy consumption data. Detecting anomalies in this data is critical for identifying abnormal energy usage patterns, preventing potential system failures, and promoting energy efficiency. In this study, we propose a statistical anomaly detection framework that integrates Exponentially Weighted Moving Average, Quantile Regression, and Adaptive Drift Detection Method to detect point anomalies in hourly energy consumption data over a year-long period. The proposed method is computationally efficient and interpretable. The findings highlight the effectiveness of this lightweight and scalable approach for real-time energy anomaly detection.