Inferring Smartphone User Demographics from Wi-Fi trace Logs: a Study of Users' Privacy Concerns

Over the recent years mobile devices have become a ubiquitous medium supporting various forms of functionality and are widely accepted for commons. However, the privacy threats along with the intimate use of smartphone has become a primary concern. A significant number of methods for persevering privacy against smartphone usage data were also proposed in recent years. The prior research mainly focuses on the privacy leakages by motion sensors, microphones, and GPS trajectories. In this paper, we report another privacy threats by analyzing a collected trace of Wi-Fi signals (referred to as Wi-Fi logs) observed by a smartphone. Such privacy leakage is neglected in the past, as Wi-Fi log data are generally considered to be less sensitive compared with GPS or microphone data. However, in this study, we show that by analyzing the Wi-Fi logs, an adversary can readily reveal many about a smartphone holder, such as occupations, moving patterns, or even user identity. To raise the concerns on this privacy leakage, we design experiments and propose a simple scheme to analyze the Wi-Fi log traces collected by recruited participants. The goal of the scheme is not to design a perfect scheme for discovering user related information but to clearly illustrate the existence and easy identification of privacy-revealing vulnerabilities in Wi-Fi trace logs. The experiment results demonstrate that the privacy can be leaked by Wi-Fi trace logs, which can be readily collected by any app requesting innocence permissions. The experiment results are alarming and may motivate the need to improve the privacy concerns by developing better privacy preserving mechanisms.