"Spatio-temporal dataset of FRMCS (5G) traffic from Railway network"

"The spatio-temporal dataset of FRMCS (5G) traffic from the railway network is a cleaned and traffic-enriched railway communication dataset designed to support research on future railway mobile communication systems. The dataset is derived from raw ETCS and railway network observations collected on the BPL high-speed railway line in eastern France from the private network of the French National Railways operator, SNCF. In this derived version, the cleaned train mobility and radio-context traces are transformed into simulation-ready spatio-temporal graph datasets for FRMCS and 5G railway studies (slicing, traffic forecasting, etc.).The dataset is organized into three generated scenarios named STGdata A, B, and C. Each scenario contains a cell-time graph table and a set of individual trains files. The graph tables aggregate railway cell occupancy, active train identifiers, emergency stop indicators, average train speed, and generated FRMCS traffic demand over critical and performance slices. The train-level files preserve real mobility and radio context used from the raw dataset collected from the real railway private network. From this real dataset, we build the graph snapshots, including cell ID, train direction, balise distance information, handover state, wideband CQI and uplink SNR estimates, subband CQI and SNR features, and application-level FRMCS traffic volumes.Application-level traffic demand in this dataset is generated from ETCS events and based on the traffic model used in the FRMCS traffic analysis conducted by the UIC (International Union of Railways). This aligns with train movement over the railway radio cells. Applications are distinguished by slice membership, enabling network slicing study. This application-level traffic part of this dataset should be interpreted as simulation-ready FRMCS traffic derived from real railway traces rather than as raw measured user-plane traffic.This dataset is suitable for research on railway network slicing, spatio-temporal graph neural networks, traffic forecasting, radio resource management, mobility-driven orchestration, handover communication analysis, emergency service prioritization, anomaly detection, and AI or ML optimization for FRMCS and 5G railway systems."