An integrated framework for reducing construction carbon emissions using real-time monitoring and econometrics.

The construction sector accounts for nearly 39% of global energy‑ and process‑related CO₂ emissions, yet its decarbonisation is hampered by the lack of real‑time, verifiable data during construction. To close this gap, we developed and validated an integrated, data‑driven framework through a case study. The framework employs a Cyber‑Physical System (CPS) with calibrated wireless sensors to stream high‑resolution operational data from construction machinery. These data were used to train a Long Short‑Term Memory (LSTM) model that predicted equipment‑level emissions with a root‑mean‑square error of 0.0196 t CO₂ and a mean absolute error of 0.015 t CO₂. A fixed‑effects panel econometric model further showed that each one‑unit rise in a regional Green Finance Index lowered construction carbon intensity by β = - 0.082 (p < 0.01). By converting granular site data into actionable insights, the framework links operational efficiency to financial reward, establishing a performance‑based paradigm for carbon management. This pathway enables policy‑makers to embed real‑time tracking into green‑finance instruments and allows practitioners to align project decisions with verified emission reductions, thereby accelerating progress toward global carbon‑neutrality goals.