Wi-Fi Crowd Monitoring: Training and Test Dataset

Abstract

This dataset provides PCAP files for developing and evaluating machine-learning methods for privacy-preserving Wi-Fi crowd sensing based on IEEE 802.11 probe requests. The PCAP files were obtained using controlled real-device captures from the MoniCrowd dataset, preserving recorded Wi-Fi transmission behaviour while introducing controlled synthetic population dynamics. Separate training and test data support the evaluation of MLP-, TCN-, and LSTM-based models for active-device population estimation.