Bitcoin users re-identification using community detection

About

Algorithm allowing the re-identification of Bitcoin users from heuristics based on the blockchain transaction history.

Usage

1. Install python dependencies from requirements file

   pip3 install -r requirements.txt

2. Setup Apache Spark, an instance will be launched by the PySpark executor

3. Setup Neo4j, the Neo4j Graph Algorithms plugin and APOC Procedures plugin and launch an instance of it

4. Create a config files in /app in the following format :

   • databaseconfig.py : PySpark and Neo4j databases configurations

   #!/usr/bin/env python
   pyspark = {
      'hdfs_path': '/path/to_json_file_or_directory',
      'memory': '2g' # Allocated memory for Spark
   }
   neo4j = {
      'uri': 'bolt://localhost:7687', # Bolt instance URI
      'user': 'neo4j',
      'password': 'password'
   }

5. Running main.py will populate the Neo4j Graph Database with the discovered user network :

   python3 main.py

Libraries

• Pyspark : https://spark.apache.org/docs/0.9.0/python-programming-guide.html

• Neo4j : https://neo4j.com/

• Neo4j Graph Algorithms plugin : https://github.com/neo4j-contrib/neo4j-graph-algorithms

• Neo4j APOC Procedures plugin : https://github.com/neo4j-contrib/neo4j-apoc-procedures

References

• Tracking bitcoin users activity using community detection on a network of weak signals - https://arxiv.org/abs/1710.08158

• Community detection in networks: A user guide - https://arxiv.org/abs/1608.00163

• The Unreasonable Effectiveness of Address Clustering - https://arxiv.org/abs/1605.06369