Reality Mining
by Imran Ali
A couple of weeks ago I was introduced to Nathan Eagle’s research on Reality Mining at MIT by eComm’s Lee Dryburgh (Nathan will be speaking at eComm 2008 in March).
Though Eagle’s work is a couple years old, it represents one of the more comprehensive studies of mobile communication and ethnography.
Supported by Nokia, the Reality Mining project has collated and mined data from the mobile handsets of 100 users and modeled various social behaviours, including conversation context, activity, proximity, location, time and relationship networks - both for individuals and aggregate groups of people.
The project’s themes have included…
- Modelling complex social systems - with applications as diverse as disease control and the social lives of freshman MIT students!
- Behaviour modelling and prediction - including generation of an automatic ‘lifelog’ of events cross referenced with the various ‘encounters’ between participants in the research. Intriguingly,researchers have attempted to model the probability of where and when people will be against their actual behaviour to determine the accuracy of predictions.
- Relationship inference - trying to automate and understand the relationship between participants based on their pattern of communication, movements and encounters.
- Social Serendipity - in the light of Dopplr’s success, the research’s investigation of serendipity seems prescient, with suggested application areas including dating, conferences and (surprisingly)Â the enterprise arena.
With the study’s raw data and client applications freely available to the world, I would expect to see handsets and services which begin to exploit the learning from this work. By embedding intelligence about my relationships and inferring behavior, could mobile communication and messaging evolve to a mixed model of explicit and overt messages with underlying exchanges of social signals, locations, moods, behavior and preference?
(Coincidentally, RW/W and MIT’s own Technology Review covered this story last week…I wonder if that’s where Lee first came across it?)
