Hello, I am Ivan Brugere, a PhD student in computer science at the University of Illinois at Chicago (with a background in data mining). I am releasing some code to help extraction of human-readable (flat text file) bitcoin transactions for analysis.
I have ported/cleaned/consolidated the code previously used by Reid and Harrigan (though not publicly available), and I think the extraction is a better methodology than the Ron and Shamir paper last autumn.
I would also be interested in soliciting any help in collecting further bitcoin data. One thing of great value would be to categorize 'nodes' (particularly high-volume nodes) based on if they are exchanges (e.g. I group and identify a few hundred addresses used by Mt. Gox), mining pools, or other. These categories would help a lot in formulating computational problems on the network.
Also any discussion of potential problems we could formulate on these datasets would be much appreciated.
Link:
http://compbio.cs.uic.edu/data/bitcoin/