EBA
The complete graph of Bitcoin's on-chain economy, built from the ground up for machine learning.
Model the financial flows and network topology of Bitcoin's on-chain economy using Bitcoin Graph.
Trace 16 years of economic evolution with a complete temporal heterogeneous graph to reveal the on-chain dynamics and behavioral patterns that static datasets obscure.
Largest Public Graph Dataset
Benchmark the scalability and performance of your models on the largest publicly available temporal graph; a heterogeneous graph of over 2.4 billion nodes and 39.7 billion time-stamped edges composed of various node and edge types.
ML-Ready Design
Bitcoin Graph abstracts away the complexities of Bitcoin's UTxO ledger, providing an intuitive representation of fund flows so you can go straight to designing and training your models.
A Complete and Accessible ML Workflow
Get started quickly with a full suite of resources. We provide the complete ETL pipeline for full data reproducibility, a customizable sampling method for creating custom subgraphs, a collection of pre-built models in ready-to-use Jupyter Notebooks, and external annotations for training or validating your models.
Anomaly Detection
Identify unusual transaction patterns, detect potentially illicit activities, or pinpoint network vulnerabilities by leveraging the graph's detailed temporal and structural information.
Entity Profiling
Analyze transaction histories and script interactions to profile different types of Bitcoin entities (e.g., exchanges, miners, services), track their evolution, and understand their economic impact within the ecosystem.
Graph ML Benchmarking
Utilize this massive, real-world graph as a challenging benchmark dataset to evaluate the performance, scalability, and effectiveness of new large-scale graph machine learning models.