Sustaining operational excellence across a geographically dispersed, high-volume retail network requires more than periodic reporting — it demands real-time data visibility, automated anomaly detection, and analytically grounded resource allocation at every node of the network simultaneously. This paper presents a production-validated framework that unifies Google BigQuery cloud data warehousing, Python-based machine learning (K-Means clustering, Isolation Forest anomaly detection, XGBoost and LSTM ensemble forecasting), and linear programming optimization into a coherent real-time operational intelligence platform deployed across The Home Depot's U.S. Regional Distribution Center and multi-store network. K-Means clustering segments 1,238 monitored store locations into four behaviorally distinct demand archetypes, enabling differentiated inventory policies and resource allocation strategies tailored to each cluster's operating characteristics. A BigQuery Medallion architecture processes more than 200 million transactional records daily with average query latencies below two seconds, replacing legacy reporting pipelines that required up to 42 minutes for equivalent computations. Tableau and Power BI dashboards — refreshing on 15-minute cycles — surface real-time KPIs, anomaly alerts, and compliance scores to operations managers across eight U.S. census regions. Over a 24-week deployment period, the framework improved network-wide inventory accuracy from 91.5% to 97.3%, increased allocation efficiency by 13.4 percentage points, raised operational compliance scores to 98.1%, and reduced the total count of active anomaly flags from 85 to 47 per monitoring cycle. The methodology advances U.S. retail supply chain resilience and is generalizable to any large-scale multi-node operational network requiring real-time decision support.