rexaro/elliptic-bitcoin-dataset

---
license: mit
task_categories:
- graph-ml
- tabular-classification
pretty_name: Elliptic Bitcoin Dataset
tags:
- bitcoin
- fraud-detection
- graph-neural-networks
- cryptocurrency
size_categories:
- 100K<n<1M
---

# Elliptic Bitcoin Dataset

## Dataset Description

This is the Elliptic Bitcoin dataset used for illicit transaction detection in cryptocurrency networks. The dataset contains Bitcoin transaction data with labeled illicit and licit transactions.

### Dataset Structure

The dataset consists of three CSV files:

1. **elliptic_txs_features.csv**: Transaction features (166 features per transaction)
- 94 local features (derived from transaction information)
- 72 aggregated features (derived from one-hop neighbors)

2. **elliptic_txs_classes.csv**: Transaction labels
- `1` = illicit (ransomware, scam, etc.)
- `2` = licit (exchanges, services, etc.)
- `unknown` = unlabeled transactions

3. **elliptic_txs_edgelist.csv**: Transaction graph edges
- Directed edges representing Bitcoin flows between transactions

### Dataset Statistics

- **Total transactions**: 203,769
- **Labeled illicit**: ~4,545 transactions
- **Labeled licit**: ~42,019 transactions
- **Unlabeled**: ~157,205 transactions
- **Time steps**: 49 (representing different time periods)

### Citation

If you use this dataset, please cite the original paper:

```bibtex
@inproceedings{weber2019anti,
title={Anti-money laundering in bitcoin: Experimenting with graph convolutional networks for financial forensics},
author={Weber, Mark and Domeniconi, Giacomo and Chen, Jie and Weidele, Daniel Karl I and Bellei, Claudio and Robinson, Tom and Leiserson, Charles E},
booktitle={Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining},
pages={1954--1964},
year={2019}
}
```

### Usage

```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("yhoma/elliptic-bitcoin-dataset")

# Access the CSV files
features_df = pd.read_csv("hf://datasets/yhoma/elliptic-bitcoin-dataset/elliptic_txs_features.csv", header=None)
classes_df = pd.read_csv("hf://datasets/yhoma/elliptic-bitcoin-dataset/elliptic_txs_classes.csv")
edges_df = pd.read_csv("hf://datasets/yhoma/elliptic-bitcoin-dataset/elliptic_txs_edgelist.csv")
```

### License

This dataset is released under the MIT License.

### Additional Information

- **Original Source**: [Elliptic Data Set](https://www.kaggle.com/ellipticco/elliptic-data-set)
- **Task**: Binary classification (illicit vs. licit transactions)
- **Suitable for**: Graph Neural Networks, LSTM, traditional ML models