A Network of Wearables

A system is presented which records audio-data from individuals, transcribes it and infers the topics of conversations and social networks from it. They are arguing that quantifying variables within speech communication is important to restructure organization charts and let groups understand themselves better. It seems to be a technology centred approach where they built it first and are trying to figure out how to use it. It is impressive that they can transcribe as much as they claim from normal speech and could have other interesting applications. Using this data for contextually-aware actions has it's own problems not easily solved by simply have more data at one's disposal. The application of this technology for understanding social networks is a trend.Since the data is recorded from real activities (in-context) it is likely to not suffer from the biases of self-reporting and memory-reliant methods.

Eagle, N., & Pentland, A. S. (2003). Social Network Computing. In A. K. Dey (Ed.), UbiComp 2003 (pp. 289–296): Springer-Verlag.

PDF