DataMatch - The Future of Data Transactions

Aditya Melkote, Maxwell Fang, Yanhao Guo, Zixin Wei
Mentor: Sheffield Nolan
📧 avmelkote@ucsd.edu, m3fang@ucsd.edu, yag007@ucsd.edu, z5wei@ucsd.edu, Sheffield.Nolan@franklintempleton.com

Abstract

Data scientists require vast amounts of data to train models. Current data marketplaces expose users to privacy risks, high costs, and security vulnerabilities.

💡 DataMatch is a decentralized, Ethereum-based data marketplace that allows users to buy and sell datasets securely, without intermediaries.

Key Features:
✅ Smart contracts for automated transactions and licensing
✅ Personally Identifiable Information (PII) Removal for privacy
✅ InterPlanetary File System (IPFS) for secure, off-chain storage
✅ Blockchain-based marketplace ensuring trustless interactions

🔗 Visit Our Live Website 🔗 GitHub Repository 🔗 Poster 🔗 Report

1. Introduction

Data is one of the most valuable assets in the digital economy. However, current online data marketplaces are:
🚨 Prone to security breaches (e.g., UnitedHealthcare data hack in 2024)
💰 Expensive, requiring hefty transaction fees
⚠️ Centralized, leading to single-point failures

Why Blockchain?
Blockchain provides a decentralized, immutable, and secure alternative for data transactions. Ethereum smart contracts enable trustless data exchanges while ensuring user privacy through cryptographic techniques.

Why IPFS?
Since blockchains struggle with scalability, we store datasets off-chain in IPFS, a peer-to-peer decentralized storage network that protects against data leaks and tampering.

2. Overview

Our project builds a secure, trustless, and scalable data marketplace by integrating:
🛡️ Ethereum-based smart contracts for transparency
📂 IPFS storage to eliminate central failure points
🔒 PII Removal to ensure privacy
💱 MetaMask integration for seamless transactions

How It Works:
1️⃣ Sellers list datasets, which are encrypted and uploaded to IPFS
2️⃣ Smart contracts handle transactions securely
3️⃣ Buyers purchase access to the dataset using MetaMask
4️⃣ Datasets remain secure & accessible via cryptographic keys

🌟 Smart contracts automate licensing, payments, and access control, making DataMatch a secure and decentralized alternative to traditional data markets.

3. Literature Review

Many decentralized marketplace proposals exist, but they struggle with usability, privacy, and scalability.

🔹 Sober et al. (2022): Blockchain-based IoT data trading reduces trust issues but lacks privacy features.
🔹 Fonseca et al. (2020): Swarm-based decentralized storage enhances scalability, but data retrieval is inefficient.
🔹 da Silva Vanin et al. (2022): Homomorphic encryption enables privacy-preserving data processing.
🔹 Li et al. (2020): Blockchain-based licensing provides automated, transparent ownership records.

Our solution combines these advancements to create a secure, scalable, and privacy-preserving marketplace.

4. Infrastructure

4.1 Core Technologies

✅ Smart Contract Development (Remix IDE, Ethereum, Amoy Testnet)
✅ Wallet Integration (MetaMask for Transactions)
✅ Data Storage (Amazon S3, Pinata, IPFS)
✅ Frontend & Backend (Next.js, Ethers.js)
✅ NLP-Based PII Removal (spaCy for Entity Detection)

4.2 How the System Works

1️⃣ Seller uploads dataset → PII detected and removed
2️⃣ Encrypted dataset stored on IPFS via Pinata
3️⃣ Smart contract records the dataset, pricing, and permissions
4️⃣ Buyer purchases dataset using MetaMask
5️⃣ Dataset remains secure, preventing unauthorized downloads

🌍 Decentralization ensures data is protected even if a single storage provider fails.

5. Methods

5.1 Smart Contract Functionality

listData(): Seller lists dataset (automatically encrypts & removes PII)
buyData(): Buyer purchases dataset (transaction confirmed via blockchain)
delistData(): Seller removes dataset from marketplace
getUserPurchased() & getCID(): Buyer accesses purchased datasets

⚙️ All transactions are immutable and recorded transparently on the Ethereum blockchain.

5.2 PII Removal Process

🔍 spaCy NLP detects:

Names
Email addresses
Phone numbers
Any other personally identifiable information

🛡️ Sanitized datasets are stored only after PII is removed.

5.3 IPFS Integration

📡 Pinata uploads datasets to IPFS, which returns a unique CID (Content Identifier).
🔗 CID is stored on the blockchain and used to retrieve files securely.

6. User Interaction

6.1 Buying Data

1️⃣ User logs into MetaMask and browses datasets.
2️⃣ Click “Buy”, triggering a smart contract transaction.
3️⃣ Dataset access granted via blockchain authentication.
4️⃣ User retrieves dataset securely using the IPFS CID.

6.2 Selling Data

1️⃣ Seller uploads dataset (PII removed, encrypted).
2️⃣ File stored in IPFS, CID recorded on blockchain.
3️⃣ Listing appears on marketplace for buyers.
4️⃣ Smart contract manages transactions securely.

🌟 Transactions are automatic, eliminating intermediaries!

7. Future Work

🚀 Planned Enhancements:
🔹 Improve market interface for better usability
🔹 Recommendation algorithm to help users find relevant datasets
🔹 Token incentives for high-quality data contributions
🔹 Advanced encryption techniques for stronger security

8. Acknowledgements

We sincerely thank:
🎓 Sheffield Nolan (Mentor)
📚 Professor Bellur
🛠️ TA Aritra Das for their guidance and support

9. References

📖 Fonseca, J., Dahal, B., & Kim, Y. (2020). Decentralized Marketplace Using Blockchain.
📖 Li, J. et al. (2020). Blockchain-Based Educational Digital Assets Management.
📖 da Silva Vanin, F. et al. (2022). Blockchain-Based Data Protection Model.
📖 Sober, M. et al. (2022). Blockchain-Based IoT Data Marketplace.