Correlating Digital Footprints for Discovering Social Connections in Crowds

This paper presents an approach for guessing the degree of "social" connection among individuals and groups moving and evolving within real environments that uses information gleaned from omnipresent surveillance and individuals digital foot prints. This task is supported by observations of individuals' behaviour within urban spaces when they are alone, and when they are part of a group (i.e., crowd). This knowledge is used as reference to predict their possibility of forming communities and how they can establish relationships with other communities in the presence of specific events (e.g., alarm, disaster). We address the challenge of combining an individual's location with a real-time graphic vision of the urban environment she is moving within, using data produced by GPS, mobile and telephone networks, and security cameras. Discovering communities within crowds leads to social graphs. Computing and processing these graphs requires computing and memory resources we therefore on our HPC infrastructure for performing tests.