#affectengines

This tumblr blog shows material that comes out of a predictive text emulator. As best I understand it takes source material then is used to write its own version of the material. This is a machine for doing something that makes strange content. It is sort of Bogostian carpentry, and sort of a unit operation come word assemblage machine.

MIT collaboration that takes your email inbox and visualises it as a social graph. You can constrain it a bit by time (week, month, year) and also curate it by scale (most connected). Not sure it does anything exciting or reveals what you don't already know, but nice example of these sorts of weighted node link clusters.