This research addresses the critical challenge of identifying energy waste in buildings, which account for approximately 40% of global energy consumption. Leveraging the Building Data Genome Project 2 dataset - an open repository containing hourly energy consumption data from 1,636 non-residential buildings - this study employs data mining techniques to detect inefficiencies and support Lean Six Sigma's waste reduction goals [1]. The methodology combines cluster analysis to identify abnormal energy patterns with correlation mining between occupancy data and HVAC usage. Implementation used Python with Matplotlib for visualization and Scipy for statistical analysis. Results revealed distinct energy consumption patterns across buildings, with clustering identifying specific consumption profiles and potential anomalies. Building_5 demonstrated significantly higher consumption peaks (exceeding 110 kWh) compared to other buildings, while correlation analysis between occupancy and HVAC usage showed a weak positive correlation (r = 0.218, p = 0.307), indicating potential inefficiencies in HVAC operation relative to actual occupancy. These findings directly connect to Lean's “7 Wastes” framework by highlighting overproduction of energy during low-occupancy periods and waiting waste from systems running unnecessarily. The research provides building managers with actionable insights for targeting energy waste reduction through data-driven decision making. This study distinguishes itself from prior works using the same dataset through: (1) a novel integration of K-means clustering with Lean Six Sigma waste framework for interpretability, (2) explicit temporal pattern analysis across working vs. non-working hours to identify occupancy-energy misalignments, and (3) a systematic anomaly detection approach linking statistical outliers to actionable waste categories. Unlike existing studies that focus on prediction or benchmarking, this work emphasizes waste detection for operational improvement.