Cities face the challenge of transforming fragmented data into actionable insights for decision-making using Local Digital Twins (LDTs). The core issue is not merely accumulating data but optimizing the flow of information to be efficient, secure, and aligned with governance for real-time applications like traffic optimization and emergency alerts. To address this, Local Digital Twins must incorporate efficient data flows using a five-layer architecture framework. This ensures purposeful, secure, resilient, and responsive flows, enabling effective decision-making. Minimizing interoperability barriers and fostering continuous feedback loops transitions digital models into full-fledged, bi-directional digital twins. This learning unit explores designing data flows using real-world case studies, emphasizing the importance of each flow layer—from sensors to application interfaces. By applying the Triple-Loop Learning model and Minimal Interoperability Mechanisms, participants map and critique data flows for better local governance and transformative public value realization.