An Open and Scalable Data Architecture for Real-Time Sustainability Intelligence

Monitoring progress toward the Sustainable Development Goals (SDGs) requires data infrastructures capable of integrating heterogeneous sustainability datasets and transforming them into actionable insights. However, many existing sustainability data systems remain fragmented, relying on isolated analytical tools and static datasets, which limit cross-domain integration and hinder the timely monitoring of sustainability indicators. This study proposes an open and scalable data architecture for sustainability intelligence that integrates modern data engineering practices with real-time analytical capabilities. The framework combines a Medallion-based data warehouse for structured data management, a star schema for multidimensional analytical querying, and an event-driven streaming pipeline that enables continuous monitoring of sustainability indicators. By integrating batch-based historical analytics with streaming-based incremental processing, the architecture supports both long-term sustainability assessment and near real-time detection of environmental anomalies. The framework is demonstrated using environmental indicators related to SDG 7 (Affordable and Clean Energy), SDG 13 (Climate Action), and SDG 15 (Life on Land), incorporating global datasets on carbon emissions, fossil fuel consumption, renewable energy generation, and forest area. Empirical evaluation through system performance benchmarking across six streaming load levels demonstrates that the architecture maintains sub-2,000 ms end-to-end latency, stable anomaly detection rates between 4.60% and 4.93%, and scalable throughput up to moderate ingestion loads, providing quantitative evidence of the framework's operational reliability for near real-time sustainability analytics. Overall, the proposed architecture provides a practical foundation for data-driven sustainability intelligence systems that support policy evaluation and evidence-based SDG monitoring.