BitConduite: Visualizing and Analyzing Activity on the Bitcoin Network
BitConduite is a visual analytics tool built to explore the activity of Bitcoin users over long-term periods.
Pseudonymous nature of transactions
With its position as the first and largest cryptocurrency, Bitcoin is attracting a plethora of different players among which are investors, governments, economists and researchers from all around the world. Transactions in Bitcoin are stored in an immutable distributed ledger that is accessible to anyone. However, this open data can be challenging to make sense of given its pseudonymous nature. Indeed, while transaction details such as the amount, the time and the sender and receiver addresses are publicly disclosed, no personal information is revealed on the identities of participants. Therefore, this abstract data does not lend itself easily to an exploratory analysis of the network actors. The increasing volume of transactions in Bitcoin is an additional challenge hampering the study of this network.
Analyzing the Bitcoin network
To visualize its billions of transactions, a team of researchers composed of Christoph Kinkeldey, Jean-Daniel Fekete and Petra Isenberg, are developing a tool to identify entities on the network from their public addresses, whether these entities are individuals or organizations. Dubbed “BitConduite,” this tool uses the network’s topology to estimate which addresses may belong to the same entity and classifies them following their activity patterns.
Since Bitcoin is used in diverse ways, from an investment asset to an illegal shopping payment system, the tool would allow us to know more about the main reasons behind Bitcoin use. Furthermore, while many argue Bitcoin is an influencing factor in world events, it is very difficult to assess such hypotheses. In 2012-2013, the financial crisis in Cyprus saw nervous cash-holders invest in Bitcoin to counter their banking system and be able to freely move their assets out of the country. Countries in crisis are no doubt raising the value of Bitcoin, but only a long-term analysis could shed light on the exact role of the cryptocurrency.
How does it work?
To make such an analysis possible, the researchers behind BitConduite extract raw data from the Bitcoin Core client and store it in a MongoDB database. This database is then tailored to visualization by using a column-oriented MonetDB database. Finally, input heuristics are applied to derive entities from pseudonymous addresses and cluster them. Analysts working with this tool can filter out entities with certain attributes, group the similar ones based on the factors of interest and visualize the number and volume of transactions of each cluster on a timeline. The development of the BitConduite tool is an ongoing work and we are very likely to see more analytics tools of this genre created to explore the use of other cryptocurrencies.