I’ll be putting some notes on last Saturday’s MobileCamp online soon, but I wanted to get my talk up for those who asked for it.

It’ll be a little tough to understand (read: impossible) without actually being at the talk, but you can check out the cool FIGletized titles if you’re into ASCII art

The main point of my talk was to try to get people thinking about data and its relationship to individuals rather than the Web 2.0 standard of mashing it up to create new hybrid datasets. I propose that users can be allowed to describe what certain types of data means to them through a set of configurable attributes and connected mathematical relationships. As data comes in to the representation, it pulses and changes dimensions as described by these relationships, but remains normalized. The result is that you have an n-dimensional parameter set that you can feed into probabilistic decision trees to make decisions on.

I exemplify this with a small conceptual game called “Can Your Life Grow a Tree”, where the user describes the data’s relationship to themselves and the data coming into the model makes decisions on how a virtual tree grows and thrives. This tree is a metaphor for decisions like “can my phone bother me now?”, “will I need to know where the Starbucks is soon?” – that kind of thing.

Thanks to everyone who attended the talk – it was awesome to have an engaging chat about the material, since I haven’t shared it with many people.

I want to give a huge mention to Rob Clewley, Computational Neuroscientist extraordinare and good friend, who developed the model with me.

You can download the powerpoint presentation here.