File size: 2,052 Bytes
249e720 40b089a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 | ---
license: cc-by-nc-sa-4.0
task_categories:
- graph-ml
tags:
- Cryptocurrency
- Bitcoin
pretty_name: EBA
---
# Overview
[EBA](https://github.com/B1AAB/EBA) interfaces with the Bitcoin network to
create a graph of the full history of on-chain transactions,
which includes the complete trading details of `>8.72B` BTC.
This temporal heterogeneous graph consists of `>2.4B` nodes and `>39.7B`
time-stamped edges spanning more than a decade,
making it an ideal resource for developing models on Bitcoin
and a large-scale benchmark for graph neural networks.
Please refer to the following paper for details:
> [Jalili, Vahid. "The Temporal Graph of Bitcoin Transactions." The Thirty-ninth Annual Conference on Neural Information Processing Systems Datasets and Benchmarks Track.](https://arxiv.org/abs/2510.20028)
We share the complete ETL pipeline and all the data it generates.
To simplify working with the pipeline and its resources,
we have split them into separate repositories:
- [ETL pipeline](eba.b1aab.ai/docs/bitcoin/etl/overview);
- [Dataset release](eba.b1aab.ai/releases/data-release/v1);
- [Sample communities](https://www.kaggle.com/datasets/vjalili/bitcoin-graph-sampled-communities);
- [_Hello-World_ model](https://github.com/B1AAB/GraphStudio/tree/main/quickstart/script_classification).
# Block Metadata
This dataset provides per-block summary statistics for the Bitcoin blockchain,
covering all blocks up to height `863 000`. The statistics are derived from four sources:
parsed from block headers, ETL logs, summarized from block content,
and generated from the chain in a post-processing step.
The goal of these stats is to provide block-level context;
either used as an independent resource
(e.g., to forecast trade volume) or to complement other datasets.
For instance, they can be combined with the Bitcoin Graph or
off-chain market indicators (like high, low, open, and close prices)
to enhance forecasting models.
[Dataset and features documentation](eba.b1aab.ai/docs/bitcoin/datasets/stats) |