Cross-Ledger Cryptocurrency Analytics Platform

GraphSense is an open source platform for analyzing cryptocurrency transaction ledgers. At the moment, it supports analysis of Bitcoin, Bitcoin Cash, Litecoin, and Zcash transactions.

Features

  • Address Clustering: GraphSense partitions the set of addresses observed in a cryptocurrency ecosystem into maximal subsets (clusters) that are likely to be controlled by the same real-world actor.

  • Network Abstractions: GraphSense computes two forms of network abstractions from the underlying blockchain: the address graph, which connects addresses via transactions, and the cluster graph, which connects address clusters.

  • Microscopic Analysis: with the GraphSense dashboard users can inspect transactions and trace currency flows by navigating along the transaction graph abstractions.

  • Macroscopic Analysis: by directly accessing the underlying, pre-computed database it is possible to map real-world phenomena (e.g., Ransomware) onto network abstractions and run subsequent analytics tasks.

  • Pre-computed statistics: GraphSense pre-computes statistics and supports interactive analysis without major delays.

  • BlockSci integration: GraphSense uses BlockSci for parsing blockchains and obtaining exchange rates.

  • Horizontal Scalability: cryptocurrency blockchains are growing and new currencies appear on the horizon. GraphSense should be future-proof, because it is built on Apache Spark and Cassandra for horizontal scalability.

Technical Architecture

GraphSense is built on scalable and distributed cluster technology and therefore requires a number of software components. They must be setup and/or executed in the following order:

Example

The following screenshot shows details about an example Bitcoin address.

screenshot

Publications

Some more technical details about GraphSense are described here; please cite as:

@inproceedings{Haslhofer:2016a,
      title={O Bitcoin Where Art Thou? Insight into Large-Scale Transaction Graphs.},
      author={Haslhofer, Bernhard and Karl, Roman and Filtz, Erwin},
      booktitle={SEMANTiCS (Posters, Demos)},
      year={2016}
  }
  

So far, GraphSense has been used for computing statistics in the following scientific papers:

Paquet-Clouston, M., Haslhofer, B., & Dupont, B. Ransomware payments in the bitcoin ecosystem. 17th annual workshop on the economics of information security (WEIS 2018). (pdf)

Filtz, E., Polleres, A., Karl, R., Haslhofer, B.: Evolution of the Bitcoin Address Graph - An Exploratory Longitudinal Study. International Data Science Conference (DSC 2017), Salzburg, Austria, 2017. (pdf)

Contributors