Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-source Multi-level Multi-token Ethereum Dataset and Benchmark Platform

Haoyuan Li, Mengxiao Zhang, Maoyuan Li, Jianzheng Li, Junyi Yang, Shuangyan Deng, Zijian Zhang, Jiamou Liu

TL;DR

3MEthTaskforce introduces a first-of-its-kind multi-source, multi-token Ethereum dataset that integrates transaction data, token metadata, global market indices, and Reddit sentiment to study user behavior, market sentiment, and token performance. It defines two benchmarks: User Behavior Prediction on a dynamic, edge-labelled temporal bipartite graph using 6 GNNs and Token Price Prediction with 19 time-series models, including LLM-enhanced sentiment features. The dataset aggregates 303 million transactions, 3,880 tokens, 35 million users, and sentiment data from 2014–2024, while adhering to FAIR data principles and privacy considerations. Experimental results show that incorporating sentiment and additional features improves predictive performance, underscoring the platform's value for risk analysis, market modeling, and DeFi research.

Abstract

This paper introduces 3MEthTaskforce (https://3meth.github.io), a multi-source, multi-level, and multi-token Ethereum dataset addressing the limitations of single-source datasets. Integrating over 300 million transaction records, 3,880 token profiles, global market indicators, and Reddit sentiment data from 2014-2024, it enables comprehensive studies on user behavior, market sentiment, and token performance. 3MEthTaskforce defines benchmarks for user behavior prediction and token price prediction tasks, using 6 dynamic graph networks and 19 time-series models to evaluate performance. Its multimodal design supports risk analysis and market fluctuation modeling, providing a valuable resource for advancing blockchain analytics and decentralized finance research.

Multi-source Multi-level Multi-token Ethereum Dataset and Benchmark Platform

TL;DR

3MEthTaskforce introduces a first-of-its-kind multi-source, multi-token Ethereum dataset that integrates transaction data, token metadata, global market indices, and Reddit sentiment to study user behavior, market sentiment, and token performance. It defines two benchmarks: User Behavior Prediction on a dynamic, edge-labelled temporal bipartite graph using 6 GNNs and Token Price Prediction with 19 time-series models, including LLM-enhanced sentiment features. The dataset aggregates 303 million transactions, 3,880 tokens, 35 million users, and sentiment data from 2014–2024, while adhering to FAIR data principles and privacy considerations. Experimental results show that incorporating sentiment and additional features improves predictive performance, underscoring the platform's value for risk analysis, market modeling, and DeFi research.

Abstract

This paper introduces 3MEthTaskforce (https://3meth.github.io), a multi-source, multi-level, and multi-token Ethereum dataset addressing the limitations of single-source datasets. Integrating over 300 million transaction records, 3,880 token profiles, global market indicators, and Reddit sentiment data from 2014-2024, it enables comprehensive studies on user behavior, market sentiment, and token performance. 3MEthTaskforce defines benchmarks for user behavior prediction and token price prediction tasks, using 6 dynamic graph networks and 19 time-series models to evaluate performance. Its multimodal design supports risk analysis and market fluctuation modeling, providing a valuable resource for advancing blockchain analytics and decentralized finance research.
Paper Structure (78 sections, 2 figures, 12 tables, 2 algorithms)

This paper contains 78 sections, 2 figures, 12 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. The 3MEthTaskforce Dataset
  4. Section 1: Token Transactions
  5. Data Collection
  6. Dataset Statistics
  7. Ethics, Privacy, IP and FAIR Data Principles
  8. FAIR.
  9. Tasks
  10. User behavior prediction
  11. Token Price Prediction
  12. User behavior prediction
  13. Baseline methods
  14. User Behavior Experiment Setup
  15. Baseline Evaluation Results
  16. ...and 63 more sections

Figures (2)

  • Figure 1: 3MEthTaskforce Data Pipeline. The third column illustrates datasets: Transaction Records (blue), Token Information (green), Global Market Indices (red), and Textual Indices (yellow).
  • Figure 2: Time-series analysis of cryptocurrency market activity and online discourse from data in this dataset. (a) displays the aggregate trading volume, indicating periods of high market activity. (b) shows the market capitalization breakdown by cryptocurrency category, revealing the relative dominance of BTC over time. (c) illustrates the temporal dynamics of submissions across various cryptocurrency subreddits. (d) presents the average positive and negative sentiment scores, revealing fluctuations in community perception of the cryptocurrency market.