ShareLikesCrowd: Mobile analytics for participatory sensing and crowd-sourcing applications

Abstract

Data and continuous data streams from mobile users/devices are becoming increasingly important for numerous applications including urban modelling, transportation, and more recently for mobile crowd-sensing to support citizen journalism and participatory sensing where sensor informatics and social networking meet. While significant efforts have focused towards the analysis of mobile user data, a key challenge that needs to be addressed in order to realize the full-potential is to address the scalability issues of real-time data collection and processing at run time. By scalability, we refer to both the challenges of data capture from a large number of users, as well as the issues of energy consumed on individual devices as a result of that capture. In this paper, we present mobile/on-board data stream mining as an effective approach to address the scalability issues of mobile data collection and run-time processing and as a significant component of mobile run-time analytics. We present experimental evaluation using the Nokia mobile data challenge open track dataset to show the significant energy and bandwidth savings that mobile data stream mining can achieve with no significant loss of useful information in this process.