Analysis of user mobility data sources for multi-user context modeling

Abstract

Finding the right data source for research is a challenge that many of us face. Although we live in times where 'Open Data' and 'Big Data' have become buzzwords, getting hold of a reasonable size and quality dataset is often hard. When it comes to user data such as mobility data, this becomes even tougher due to privacy-related concerns. This paper briefly explains our research in the area of multi-user context modeling and presents some criteria that we believe are important while selecting a dataset for testing different approaches in this domain. To find the right dataset, some relevant publicly available human mobility datasets are examined using these criteria. The following are the datasets that have been analyzed: Microsoft Research GeoLife Trajectory Dataset, Tracking Delft I Pedestrian Trajectory Dataset, MIT Media Lab Reality Mining Dataset and LifeMap Dataset. Besides these, some other useful data sources for researchers have been cited.