Spotsensing

Social Networking Sites (SNS) have an unprecedented ability to capture Human activity, including information about the specific physical settings in which those activities are taking place. This represents a major potential for uncovering, on a large scale, new knowledge about aggregate behaviors in the use of places. In this paper, we explore the concept of social web sensor, as a systematic data collection process that can be virtually attached to a particular location to retrieve location-based information from social network sites. This process is completely based on geographically scoped queries to SNS APIs and does not depend on real physical sensors. The objective of this study is mainly to assess the viability of this concept and uncover the potential and limitations of this approach as a reality mining tool for urban environments. We have created an initial implementation and conducted the respective evaluation through the deployment of a number of sensors in the city of London and the analysis of the respective results.