Automated BIM-linked analytics for construction project control: a proof-of-concept framework with headset-accessible delivery

Construction project-control data spans modelling, scheduling, and cost systems, making it difficult to trace performance indicators to the physical work that produces them. This paper specifies and demonstrates a five-phase framework for automated, model-linked project control. Phase 1 defines a relational schema. Phase 2 links BIM (Revit) elements to Primavera P6 activities and cost records via Activity ID and Location keys, applying pre-publication validation rules classified by severity. Phase 3 computes earned-value indicators within a Power BI semantic model using DAX. Phase 4 automates refreshes via Power Automate. Phase 5 delivers results as a headset-accessible browser page featuring a Speckle model stream alongside the dashboard. Developed following design-science principles, the framework was demonstrated on an active commercial project (data date: 15 February 2025) with a budget at completion of ∼ EGP 125.3 million within a ∼ EGP 136 million contract, a schedule performance index of 0.87, and a cost performance index of 0.90. The paper details relational grain, keys, cardinalities, fifteen data-quality validation rules, DAX what-if formulations, and one-hour refresh constraints. Delivering a 2D headset browser page that provides headset-accessible visualization over immersive analytics and contributing a reproducible integration specification; evaluating accuracy, latency, and user performance remains further work.