A casualty-estimation framework has been proposed that incorporates building-scale, time-varying occupancy data (considering factors such as size, use, and time of day) into earthquake loss modeling. Information from Seumter building records is combined with KOSIS population data, and HAZUS modules are used to estimate both the baseline population and dynamic occupancy at the building level. Case studies have shown a close alignment with observed data, with no significant model flaws, indicating the framework’s operational readiness. This approach moves beyond broad administrative totals to provide micro-spatial resolution suitable for Korea’s rapid seismic attenuation and localized damage patterns. It enables accurate, time-sensitive casualty estimates. The framework is designed to be scalable to include additional data sources, such as mobility, transportation, and activity patterns. It supports effective evacuation and shelter planning, surge capacity management, and prioritization of retrofitting, leading to more efficient resource allocation. Furthermore, the framework provides a consistent method to integrate future data streams and quantify uncertainty without disrupting the core workflow.