NFTDisk: Visual Detection of Wash Trading in NFT Markets

Title of the Paper

NFTDisk: Visual Detection of Wash Trading in NFT Markets

Paper Information

• Subject Area: Visual analysis and detection of wash trading in NFT markets
• Keywords: Non-Fungible Tokens (NFT), wash trading, visual analysis, fintech, blockchain, data visualization, transaction networks, user behavior analysis

Research Background and Problem

• Problem or Challenge: With the growing popularity of NFT markets, various fraudulent activities have also increased. Among these, wash trading has become one of the most common forms of fraud, fabricating transaction volumes to mislead investors. Traditional automated detection algorithms fail to identify all wash trading activities due to the complexity of trading patterns, often requiring manual inspection.
• Significance: Wash trading severely impacts market transparency, misleads investors, and can result in significant financial losses. Understanding and analyzing wash trading can help protect investors and regulate the market.
• Research Motivation and Related Work:
  • Existing research has proposed network-structure-based algorithms for detecting wash trading, but these methods face limitations in detection and false positives due to the evolving strategies of wash trading.
  • Visualization analysis techniques have been used to understand blockchain data, but existing solutions are mostly designed for traditional cryptocurrencies like Bitcoin and fail to address the unique characteristics of NFT transactions.

Solution

• Main Approach: An interactive visualization tool called NFTDisk is proposed for detecting and analyzing wash trading in NFT markets.
• Innovations:
  1. Designed a radial visualization module based on a disk analogy for a quick overview of NFT transactions and identification of wash trading patterns.
  2. Integrated a flow visualization module to display detailed multi-level transaction flow information.
  3. Proposed a new address reordering strategy to optimize the layout of high-frequency trading addresses, reducing visual clutter.
  4. Provided rich interactive features, including multi-level filtering and analysis based on time or addresses.
• Implementation Steps:
  1. Present an overall view of NFT transactions in the disk module, marking suspicious addresses based on transaction volume and patterns.
  2. Refine the analysis through the flow module, displaying transaction paths of suspicious address groups and changes in NFT holdings over time.
  3. Support investors in filtering irrelevant information and focusing on wash trading behaviors through user interaction features (e.g., brushing, zooming).

Research Outcomes

• Specific Achievements:
  1. Developed a complete toolchain for visualizing and detecting wash trading behaviors in NFT transactions.
  2. Validated the effectiveness and practicality of NFTDisk in real-world investment scenarios through interviews with 14 NFT investors and case studies (e.g., Meebits and Loot transaction collections).
  3. Captured multiple wash trading patterns, such as complex bot trading groups and high-frequency trading behaviors driven by transaction rewards.
• Advantages:
  • Compared to existing algorithm-only methods, NFTDisk explicitly visualizes trading patterns and helps users intuitively understand complex wash trading behaviors.
  • Its modular design supports multi-level analysis, enabling intuitive exploration from global transactions to individual NFTs, and from address groups to specific paths.
• Experiments and Evaluation:
  • Case studies demonstrate that NFTDisk-based analysis significantly reduces the time required to identify wash trading.
  • Feedback from user interviews shows that investors particularly appreciate its ability to filter suspicious transactions and perform multi-level analysis, with overall high ratings.
• Limitations and Future Directions:
  1. For NFT collections with a large number of addresses or high transaction volumes, the current screen space may be insufficient for effective presentation.
  2. Existing data is primarily based on on-chain records, lacking tracking of more complex market behaviors (e.g., batch trading tools).
  3. Future work proposes incorporating automated algorithms to assist with initial filtering and integrating account balances and other off-chain data to enhance comprehensive analysis capabilities.

Through this research, NFTDisk not only addresses key visualization challenges in wash trading detection but also provides an intuitive and practical tool for in-depth investigation of fraudulent activities in NFT markets. It offers significant contributions to the fields of fintech and blockchain data analysis.