[Air-L] mailman exploration tool project (and an air-l network visualization)

[Bernhard Rieder]

Hello,

I'm currently preparing classes for the new semester and one of the 
things I'll do with my students (multimedia) is working on digital 
methods and the visualization of large amounts of data.

One of the things we're going to work on is a tool for exploring mailman 
archives (they're pretty much standard and easy to parse) and I would 
like to know whether anybody has an ongoing research project on (a) 
mailinglist(s) and some data we could play with.

I've written some quick code and here 
(http://software.rieder.fr/stuff/airl_undirected.png) you'll find a 
visualization of the air-l archives (22K mails, 3K posters) that shows a 
social graph where replying to a mail constitutes a relation (min 2 
interactions, link strength taken into account). Only posters with 10+ 
mails taken into account. Node size is no of posts, node color is degree.

A working toolkit should provide basic statistical/topological analysis, 
statistical exploration of word frequencies, topic clustering, data 
export, anonymization, and a way to visualize network structures over time.

This stuff needs some serious domain knowledge to be even remotely 
interesting (beyond the "look it's a map") and I'd love to work with 
somebody who has an actual research project and wouldn't mind giving 
occasional feedback.

best,
B.

-- 
Bernhard Rieder
Laboratoire Paragraphe
Université de Paris VIII

bernhard.rieder at univ-paris8.fr
http://bernhard.rieder.fr
http://thepoliticsofsystems.net