Speaker – Nathan Eagle
email - i forgot :)
Nathan is a research assistant at MIT Labs, currently in Kenya where he works with KEMRI in Kilifi and with University students in UoN (why UoN only?) to develop mobile applications.
He gave a talk on his Phd research whose aim was to use data obtained from people's mobile phones to demonstrate social patterns and predict their social behaviours. Basically, social studies based on mobile phone data. I hope i am right on that one.
He developed this cool application that runs in the background of a mobile phone collecting all sorts of statistics.. e.g, calls made, messages received, which cell towers the mobile phone is, e.t.c. The data was sent monthly via GPRS to a server. He actually got kicked out by AT&T coz of the large amount of data he was transferring coz they saw it as a Denial of Service attack.
He also had stations throughout the campus with devices that continually scanned and recorded all bluetooth devices around the station.
Next step - analysis of the data, and he actually showed that patterns were repeated every 24 and every 168 hours. He could do all kinds of cool things with the data...even predict who is a friend to a subject.. with relatively good accuracy. He used some mathematical models that were quite frankly, scaring. Some very complicated and huge formulas.
The discussion then turned to why don't the mobile phone companies provide the data to the subscribers? Kina safcom and celtel actually have all this data. Suggestions were made as to how the data can be useful in real life, e.g. for observation of traffic patterns, or even auto-diaries. Mobile phone applications were also discussed, with the question of whether to use J2ME or Python for the applications arising. I didnt even know that python can be used for mobile applications.
He also has data about all calls made in UK for Aug. 2005. He is trying to make sense of it by making some advanced mathematical models... Good Luck!
Nathan - can you share the presentation online?