Hugging Face's logo

Datasets:
MLNTeam-Unical
/
NFT-70M_transactions

Tasks:
Time Series Forecasting
Text Classification
Feature Extraction
Modalities:
Tabular
Text
Formats:
parquet
Languages:
English
Size:
10M - 100M
Tags:
Non-fungible Tokens
Crypto
Web3
Art
Multimodal Learning
DOI:
Libraries:
Datasets
Dask
License:
Dataset card Viewer Files Community
3
NFT-70M_transactions / README.md
MLNTeam-Unical's picture
MLNTeam-Unical
Update README.md
63c2408
preview code
|
raw
history blame contribute delete
16.3 kB
metadata
dataset_info:
  features:
    - name: num_sales
      dtype: int64
    - name: fees_seller
      dtype: float64
    - name: fees_opensea
      dtype: float64
    - name: fees_seller_usd
      dtype: float64
    - name: fees_opensea_usd
      dtype: float64
    - name: tx_timestamp
      dtype: string
    - name: price
      dtype: float64
    - name: gain
      dtype: float64
    - name: usd_price
      dtype: float64
    - name: usd_gain
      dtype: float64
    - name: token
      dtype: string
    - name: to_eth
      dtype: float64
    - name: to_usd
      dtype: float64
    - name: created_date
      dtype: string
    - name: chain
      dtype: string
    - name: token_type
      dtype: string
    - name: asset_contract_type
      dtype: string
    - name: asset_type
      dtype: string
    - name: payout_collection_address
      dtype: int64
    - name: from_account
      dtype: int64
    - name: to_account
      dtype: int64
    - name: seller_account
      dtype: int64
    - name: winner_account
      dtype: int64
    - name: contract_address
      dtype: int64
    - name: nft_image
      dtype: int64
    - name: collection_image
      dtype: int64
    - name: token_id
      dtype: int64
    - name: nft_name
      dtype: int64
    - name: nft_description
      dtype: int64
    - name: collection_name
      dtype: int64
    - name: collection_description
      dtype: int64
  splits:
    - name: train
      num_bytes: 21291348001
      num_examples: 70972143
  download_size: 6633664673
  dataset_size: 21291348001
size_categories:
  - 10M<n<100M
license: cc-by-nc-4.0
task_categories:
  - time-series-forecasting
  - text-classification
  - feature-extraction
  - text-generation
  - zero-shot-classification
  - text2text-generation
  - sentence-similarity
  - image-classification
  - image-to-text
  - text-to-image
  - text-retrieval
language:
  - en
tags:
  - Non-fungible Tokens
  - Crypto
  - Web3
  - Art
  - Multimodal Learning
pretty_name: NFT-70M_transactions

Dataset Card for "NFT-70M_transactions"

Dataset summary

The NFT-70M_transactions dataset is the largest and most up-to-date collection of Non-Fungible Tokens (NFT) transactions between 2021 and 2023 sourced from OpenSea, the leading trading platform in the Web3 ecosystem. With more than 70M transactions enriched with metadata, this dataset is conceived to support a wide range of tasks, ranging from sequential and transactional data processing/analysis to graph-based modeling of the complex relationships between traders. Besides, the availability of textual and image contents further amplifies the modeling capabilities and usage opportunities of this dataset, making it a unique and comprehensive multimodal source of information for delving into the NFT landscape.

This dataset can serve as a benchmark for various innovative and impactful tasks within the crypto landscape, such as projecting NFT prices or detecting fraudolent and wash trading activities. Furthermore, the multimodal nature of the dataset fosters the development of classification models, as well as textual and visual generative models.

Data anonymization

We point out that the collected NFT transactions and metadata from OpenSea are publicly distributed on blockchain. For our purposes of re-distribution, we are also committed to ensure non-disclosure of information that might lead to identifying the NFT creators, in order to be compliant with privacy-preserving requirements and to avoid violation of data protection regulations and of property rights. In this respect, we carried out three actions:

  • Values of all variables describing non-sensitive information were kept in their original form;
  • Values of all variables describing sensitive information were anonymized, in a one-way, non-revertible mode;
  • URLs of image data and textual contents (i.e., NFT images and their descriptions) were replaced by identifiers to numerical vectors that represent an encrypted representation (i.e., embeddings) of the image/text contents obtained via neural network models. Such embeddings are eventually provided in place of their original image and text data, and can be found in the NFT-70M_image and NFT-70M_text supplementary datasets, respectively.

Data Fields

Variable Type Description Processing Notes
token_id String The id of the NFT — this value is unique within the same collection Anonymized Original values were replaced by hash-codes
num_sales Integer A progressive integer indicating the number of successful transactions involving the NFT up to the current timestamp (cf. tx_timestamp) Original Not sensitive variable
nft_name Vector ID The name of the NFT Anonymized Original values were encrypted via neural textual embedding
nft_description Vector ID The description of the NFT as provided by the creator Anonymized Original values were encrypted via neural textual embedding
nft_image Vector ID The ID for accessing the NFT image vector Anonymized Original values were encrypted via neural visual embedding
collection_name Vector ID The ID for accessing the Collection name vector Anonymized Original values were encrypted via neural textual embedding
collection_description Vector ID The ID for accessing the Collection description vector Anonymized Original values were encrypted via neural textual embedding
collection_image Vector ID The ID for accessing the Collection image vector Anonymized Original values were encrypted via neural visual embedding
fees_seller Float The absolute amount of fees the seller has gained from this transaction expressed in token Original Not sensitive variable
fees_opensea Float The absolute amount of fees OpenSea has gained from this transaction expressed in token Original Not sensitive variable
fees_seller_usd Float The absolute amount of fees the seller has gained from this transaction expressed in US dollars (USD) Original Not sensitive variable
fees_opensea_usd Float The absolute amount of fees OpenSea has gained from this transaction expressed in US dollars (USD) Original Not sensitive variable
payout_collection_address String The wallet address where seller fees are deposited Anonymized Original values were replaced by hash-codes
tx_timestamp String Timestamp of the transaction expressed in yyyy-mm-ddTHH:MM:SS Original Not sensitive variable
price Float The price of the transaction expressed in token Original Not sensitive variable
gain Float The gain after fees (i.e., gain = price - fees_opensea * price - fees_seller * price) Original Not sensitive variable
usd_price Float The price of the transaction expressed in US dollars (USD) Original Not sensitive variable
usd_gain Float The difference between the price and the fees expressed in US dollars (USD) Original Not sensitive variable
token Categorical The token type used to pay the transaction Original Not sensitive variable
to_eth Float The conversion rate to convert tokens into Ethereum at the current timestamp, such that eth = price * to_eth Original Not sensitive variable
to_usd Float The conversion rate to convert tokens into US dollars (USD) at the current timestamp, such that usd = price * to_usd Original Not sensitive variable
from_account String The address that sends the payment (i.e., winner/buyer) Anonymized Original values were replaced by hash-codes
to_account String The address that receives the payment (it often corresponds to the contract linked to the asset) Anonymized Original values were replaced by hash-codes
seller_account String The address of the NFT seller Anonymized Original values were replaced by hash-codes
winner_account String The address of the NFT buyer Anonymized Original values were replaced by hash-codes
contract_address String The contract address on the blockchain Anonymized Original values were replaced by hash-codes
created_date Timestamp The date of creation of the contract Original Not sensitive variable
chain Categorical The blockchain where the transaction occurs Original Not sensitive variable
token_type Categorical The schema of the token, i.e., ERC721 or ERC1155 Original Not sensitive variable
asset_contract_type Categorical The asset typology, i.e., non-fungible or semi-fungible Original Not sensitive variable
asset_type Categorical Whether the asset was involved in a simple or bundle transaction Original Not sensitive variable

How to use

Data provided within this repository can be straightforwardly loaded via the datasets library as follows:

from datasets import load_dataset
dataset = load_dataset("MLNTeam-Unical/NFT-70M_transactions")

Complementary data involving textual and visual embeddings can be integrated as follows:

from datasets import load_dataset
import numpy as np


transactions_dataset=load_dataset("MLNTeam-Unical/NFT-70M_transactions")
image_dataset=load_dataset("MLNTeam-Unical/NFT-70M_image")
text_dataset=load_dataset("MLNTeam-Unical/NFT-70M_text")


# Mapping from image_id to the row_index within the image dataset
image_id2row_index={int(id):k for k,id in enumerate(image_dataset["train"]["id"])}

# Mapping from text_id to row_index within the text dataset
text_id2row_index={int(id):k for k,id in enumerate(text_dataset["train"]["id"])}


def get_image_embedding(image_id,image_id2row_index,image_dataset):
  # If the mapping contains the image, the embedding exists
  idx_emb=image_id2row_index.get(int(image_id),None)
  
  if idx_emb:
    # If the embedding exists, return it
    return np.array(image_dataset["train"].select([idx_emb])["emb"][0])
  else:
    return None
  
def get_text_embedding(text_id,text_id2row_index,text_dataset):
  # If the mapping contains the text, the embedding exists
  idx_emb=text_id2row_index.get(int(text_id),None)

  if idx_emb:
    # If the embedding exists, return it
    return np.array(text_dataset["train"].select([idx_emb])["emb"][0])
  else:
    return None

### USAGE EXAMPLE ###

# Select transaction_id
transaction_id=120

# Get the image_id (e.g., collection_image or nft_image)
id_image=transactions_dataset["train"].select([transaction_id])["collection_image"][0]

# Get the image
image_embedding=get_image_embedding(id_image,image_id2row_index,image_dataset)

# Get the text_id
id_text=transactions_dataset["train"].select([transaction_id])["collection_description"][0]

# Get the text
text_embedding=get_text_embedding(id_text,text_id2row_index,text_dataset)

Ethical use of data and informed consent

This data repository is made available for research and informational purposes only.

Any finding that might be drawn from the data provided within this repository should be intended to support decision-making regarding actions made on NFTs, and not to replace the human specialists.

The authors are not responsible for any issues related to trading failures based on the data provided within this repository.

Terms of Usage

Please cite the following papers in any research product whose findings are based on the data provided within this repository:

  • L. La Cava, D. Costa, A. Tagarelli: SONAR: Web-based Tool for Multimodal Exploration of Non-Fungible Token Inspiration Networks. In: Proc. ACM SIGIR 2023. Taipei, Taiwan, July 23-27 2023. DOI: https://doi.org/10.1145/3539618.3591821
  • L. La Cava, D. Costa, A. Tagarelli: Visually Wired NFTs: Exploring the Role of Inspiration in Non-Fungible Tokens. CoRR abs/2303.17031 (2023). DOI: https://doi.org/10.48550/arXiv.2303.17031
  • D. Costa, L. La Cava, A. Tagarelli: Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction. In: Proc. ACM WebConf 2023, pp. 1875-1885. Austin, TX, USA, 30 April 2023 – 4 May 2023. DOI: https://doi.org/10.1145/3543507.3583520

Data within this repository were fetched using the REST APIs provided by OpenSea. You should also acknowledge OpenSea API.

Liability statement

The authors hereby declare that they are not responsible for any harmful or objectionable content that may be contained within the data provided within this repository. Users of the dataset are expected to exercise due diligence and responsibility when using the data, including but not limited to: (i) Content Review: Users should review the dataset's contents carefully and assess its suitability for their intended purposes; (ii) Compliance: Users are responsible for ensuring that their use of the dataset complies with all applicable laws, regulations, and ethical standards; (iii) Data Processing: Users may need to apply data preprocessing, filtering, or other techniques to remove or address any objectionable or harmful content as needed.

The authors of this dataset disclaim any liability for the accuracy, completeness, or suitability of the data and shall not be held responsible for any consequences resulting from the use or misuse of the dataset.

By accessing and using this dataset, users acknowledge and accept this disclaimer.