Pumpfun_v2_dataset / README.md
Slinky21's picture
Update README.md
c4ee64e verified
|
Raw History Blame Contribute Delete
2.15 kB
metadata
license: cc-by-4.0
language:
  - en
tags:
  - crypto
  - solana
  - pumpfun
  - memecoins
  - pump.fun
  - research
  - q-fin
pretty_name: Pumpfun_dataset
size_categories:
  - 10B<n<100B

Pumpfun Dataset

License & Usage Terms

  • License: Creative Commons Attribution 4.0 International (CC-BY-4.0)
  • Permitted Uses: Internal research, commercial/non-commercial local analysis, and modeling are fully permitted.
  • Attribution Requirement: You are free to share, transform, build upon, or analyze this dataset ONLY IF you give appropriate credit to the author, provide a link to the license, and indicate if changes were made.
  • Credit Format: Please attribute this dataset as:

    Data provided by [Slink21]. Pumpfun Solana Dataset (2026).


Data Provenance & Methodology

1. Data Source

  • Origin: This dataset is directly ingested from our own dedicated, custom Solana RPC node and Geyser gRPC stream monitoring live program logs and instruction data on-chain.
  • No 3rd Party APIs: No middleman APIs or third-party indexers (e.g., Birdeye, Helius API) were used, ensuring full raw chain fidelity and low latency.
  • Rights & Licensing: As on-chain transaction data broadcasted to the public Solana blockchain is public Domain / uncopyrightable raw factual data, we hold full rights to structure, format, compile, and license this structured compilation for public research under CC-BY-4.0.

2. Field Definitions & Data Generation

  • source_topic: Generated by our indexing pipeline to categorize the specific stream event or program instruction type (e.g., pump_fun.create_token, pump_fun.buy, pump_fun.sell, or pump_fun.complete).
  • ingested_at: The UTC ISO-8601 timestamp generated at the exact millisecond our custom parser parsed and validated the transaction event off our Geyser stream buffer.

Data Retention & Reproducibility

  • Local Retention: Yes, you are explicitly allowed to retain local, unmodified copies of all parquet/JSON/CSV files indefinitely for research reproducibility, backtesting, and auditing purposes.