Bitcoin Graph v1

This data release provides the full output of the ETL pipeline, covering the Bitcoin blockchain up to block height 863 000. The release includes the complete Bitcoin Graph, consisting of >2.4B nodes, >39.7B edges, provided in both Neo4j-compatible TSV and database dump formats, alongside pre-sampled communities and per-block summary statistics.

The release contains the following resources:

---

Bitcoin Graph

Use Cases

• Developing specialized cryptocurrency ML models;
• Sampling application-specific communities;
• Exploring the graph.

Downloads

• Download Neo4j format (1.17TB) from:

  • Google Drive
  • AWS: s3://bitcoin-graph/v1/data_to_import_neo4j/ (Walkthrough)

• Download Neo4j database dump (731GB) from:

  • Google Drive
  • AWS: s3://bitcoin-graph/v1/neo4j_db_dump/ (Walkthrough)

Schema & Specs

• Neo4j format (TSV) schema

---

Sampled Communities

Use Cases

• A quick introduction and exploring the dataset.

Downloads

• Download from Kaggle
• Use in a Hello-World script classification model

---

Block Summary Statistics

Use Cases

• Longitudinal analysis of on-chain dynamics and trends.
• Combine on-chain block statistics with external data (e.g., market trading values for use in market forecasting models).

Downloads

• Download from Kaggle
• Documentation

Schema & Specs

• Block Statistics schema

---

Dataset Spec

• Covers all blocks on the Bitcoin Mainnet up to block height 863 000.

• Node count: 2 405 686 021 (>1.3B Script nodes, 1.08B Tx nodes, 863k Block nodes, 1 Coinbase node).

• Edge count: 39 724 622 327 (>13.2B Redeems, >9.6B Credits, >8.85B Transfers, >4.3B Fee, >3.6B Confirms, >3.3M Mints).

• The following transactions are excluded:

  • Transactions where the number of non-zero BTC input and output scripts is more than 20 each are excluded.
  • Transactions and UTxO with 0 BTC are excluded.