Fraudcoins/crypto-holder-concentration

FraudCoins on-chain holder-concentration dataset

Daily on-chain measurements of top-10 holder concentration for cryptocurrency tokens, plus contract-permission data — published under CC BY 4.0 and archived with a DOI.

DOI: 10.5281/zenodo.22494421 · Docs: <https://fraudcoins.com/data/> · Live API: <https://fraudcoins.com/dataset/v1/> (no key, CORS enabled)

What makes it different

A raw "top 10 holders hold X%" figure counts whatever the ten largest addresses happen to be — routinely an exchange's omnibus wallet, a liquidity pool, a bridge contract, a locked treasury or a burn address. Those are custodial or pooled balances, not one person deciding whether to sell.

This dataset classifies each of the ten largest holders first, publishes the share held by private wallets only, and — uniquely — names every wallet it set aside, with its classification and stake, in excluded-holders.json. The correction is auditable on a block explorer rather than asserted.

Files

Reading it correctly

1. 1.`null` means no measurement — never a clean result. An unreported contract permission is unknown, not absent.
2. 2.A flat interval in `history.json` means no observed change, not a daily re-measurement. Consecutive identical readings are not re-emitted.
3. 3.The date field differs per file: sampledAt, asOf, measuredAt.
4. 4.Flag thresholds are per tier, defined in concentration.json's tiers array, and a flag requires at least 50 holder records.
5. 5.`alerts.json` entries are mostly `first-measured`, not crossed — the asset was already above the threshold when first measured. Reading one as "this project newly deteriorated" would be wrong.

Limitations

• —Holder classification relies on public address labels; an unlabelled custodial wallet is counted as private, so every correction is a floor.
• —Excluded supply is not automatically harmless — treasury and vesting holdings can reach the market on unlock.
• —Contract data is third-party static analysis of the deployed contract. Trade-simulation verdicts (honeypot, cannot-sell-all, taxes) are deliberately excluded: they false-positive on legitimate, freely-traded assets.
• —Risk scores and flags are automated, opinion-based assessments derived from public market data — not statements of fact, financial advice, or allegations of illegal conduct against any project or person.

Citation

@dataset{fraudcoins_dataset,
  title  = {FraudCoins open dataset: on-chain holder concentration and contract permissions},
  author = {{FraudCoins.com}},
  year   = {2026},
  doi    = {10.5281/zenodo.22494421},
  url    = {https://doi.org/10.5281/zenodo.22494421}
}