The rapid expansion of IoT deployments has intensified the need for efficient real-time sensor data aggregation. Conventional cloud-centric methods suffer from high latency and bandwidth limitations, making them unsuitable for latency-sensitive applications. To address these challenges, we developed a structured sensor data aggregation architecture comprising edge devices, fog nodes, and cloud infrastructure. The system was evaluated using synthetic and public datasets across sensor networks ranging from 50 to 1600 nodes. Edge device processing maintained latency below 200 ms (129.73–196.85 ms), while fog node aggregation reduced bandwidth usage by up to 85%. Overall, the architecture achieved a 95% reduction in bandwidth consumption compared with cloud-only solutions. Accuracy declined from 94.23 to 80% as sensor density increased, and throughput dropped by approximately 90% (from 1805.01 to 169.24 events/s). Energy efficiency decreased from 91.91 to 20.57 arb. unit. The integrated preprocessing pipeline—combining wavelet denoising, spatiotemporal imputation, and multi-method outlier detection—improved accuracy by 22–32%. The architecture demonstrated adaptability across healthcare, smart cities, and industrial control systems, supporting sub-second response times and scalable deployment. These results validate the architecture’s viability for real-time IoT applications, while highlighting the need for further optimization in dynamic environments and resource-constrained edge devices.