Code, Datasets and SPSN'11

If you would like to generate a text-file describing the entire list of Bitcoin transactions from your local data directory, we have forked Gavin Andresen's bitcointools project to include an "--all-transactions" option.  It produces a tab-delimited text-file where each line corresponds to one input or output of a transaction (in/out) and has one of the following formats:

• in, hash, coinbase
• in, hash, prev_hash, prev_index, pubkey
• out, hash, index, pubkey, value

For the analysis in our previous blog post, we constructed three networks from this text-file: the transaction network, the public-key network and the user network. They are essentially directed graphs with attributes. Their construction is described in the preprint on arXiv. The networks were constructed on 13th July 2011.
To help others verify our work, and allow further academic study of the Bitcoin networks, we are making these networks, generated from the public transaction history, available for download:

• The Transaction Network (transaction_vertices_2011-07-13.txt, transaction_edges_2011-07-13.txt)
• The Public-Key Network (public_key_vertices_2011-07-13.txt, public_key_edges_2011-07-13.txt)
• The User Network (user_vertices_2011-07-13.txt, user_edges_2011-07-13.txt)

In each case, the first two columns in an edge list reference line numbers in the corresponding vertex list. For example, the first entry in user_edges_2011-07-13.txt ("1 - 5994 - 8.94 - 2011-07-04-09-05-56") indicates that the user represented by line number 1 of user_vertices_2011-07-13.txt sent the user represented by line number 5944, 8.94BTC on the 4th July 2011.

Finally, we will be presenting a paper describing our work at the First International Workshop on Security and Privacy in Social Networks 2011 (SPSN'11) next month.