An Urban Multi-Operator QoE-Aware Dataset for Cellular Networks in Dense Environments

Résumé fourni par la source

This dataset contains 30,925 labelled and cleaned records collected from a dense 2 km² urban area surrounding Sunway University, Selangor, Malaysia. Using the GNetTrack Pro mobile application and Samsung S21 Ultra devices, the data spans three anonymized commercial mobile network operators and includes both 4G and 5G technologies. The dataset captures radio signal quality metrics (RSRP, RSRQ, SNR, etc.), geospatial information, mobility patterns (walking vs. driving), and application-specific traffic scenarios (HTTP, FTP, 1080p Video Streaming). A total of 132 physical cell sites were validated via OpenCellID and field inspections. The dataset is released in CSV format and includes Python scripts for data preprocessing and basic visualization. This makes it a valuable resource for machine learning tasks like signal metric regression, handover optimization, and QoE modeling in heterogeneous and simulation of high-density urban networks. Key features: Real-world 5G/4G measurements Multi-operator and multi-mobility modes Traffic-aware profiling Empirical validation of base station locations Ready for ML/DL use cases **Note: The updated processed_data_cleaned.v2.csv dataset uses 3 mobility modes (BRT, Shuttle and Walk) in addition to the oversimplified modes two modes used previously.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.