NoiseBay: A Real-World Study on Transparent Data Collection

Abstract

In applications where data is collected with the help of personal mobile devices, very often, from the user’s point of view, opaque and partly uncontrollable processes are running in the background of devices. In this paper, we show the advantages of an alternative participant-controlled transparent data collection approach. The paper combines a detailed experimental real world study with a best-practice report. We study the discrepancy between the transparency in the data collection process and the quality of the data collected in the context of mobile crowdsensing (MCS), a paradigm which leverages sensing data from the mobile devices of private individuals. We focus on applications where environmental data is collected and private user data in itself should not have any additional benefit. We treat the concrete example of MCS of tempo-spatial data for the creation of a thematic map of noise levels. We present a lightweight transparent online scheduling approach of opt-in requests for data collection for the users. Within the framework of a real world study, we show that our approach is competitive and results in an improved workload balance among users. We also present data on the responsiveness of users to requests.