SGVis: Analysis of Data from Mass Participation Ubicomp Trials

The recent rise in popularity of ‘app store’ markets on a number of different mobile platforms has provided a means for researchers to run worldwide trials of ubiquitous computing (ubicomp) applications with very large numbers of users. This opportunity raises challenges, however, as more traditional methods of running trials and gathering data for analysis might be infeasible or fail to scale up to a large, globally-spread user base. SGVis is a data analysis tool designed to aid ubicomp researchers in conducting trials in this manner. This paper discusses the difficulties involved in running large scale trials, explaining how these led to recommendations on what data researchers should log, and to design choices made in SGVis. The authors outline several methods of use and why they help with challenges raised by large scale research. A means of categorising users is also described that could aid in data analysis and management of a trial with very large numbers of participants. SGVis has been used in evaluating several mass-participation trials, involving tens of thousands of users, and several use cases are described that demonstrate its utility.