README.md

---
title: Chainticks
emoji: 📊
colorFrom: blue
colorTo: green
sdk: static
pinned: false
---

# Chainticks

Chainticks publishes agent-friendly crypto, perp, on-chain, and market-structure datasets for quantitative research.

The public Hugging Face datasets are designed to be easy to inspect from Python, DuckDB, Polars, pandas, and autonomous research agents. Each dataset includes:

- partitioned Parquet files
- a machine-readable `_schema.json`
- an append-only `_manifest.json`
- a `LATEST_DATE.txt` pointer
- source-kind provenance fields

## Public Datasets

- [Perp Data](https://huggingface.co/datasets/Chainticks/perp-data): Hyperliquid chain/archive-derived funding, trades, markets, open interest, and liquidations.
- [Stablecoin Flows](https://huggingface.co/datasets/Chainticks/stablecoin-flows): public ERC-20 stablecoin mint, burn, and bridge-style flow rows.
- [DEX Swaps](https://huggingface.co/datasets/Chainticks/dex-swaps): normalized public EVM swap events.
- [DeFi Liquidations](https://huggingface.co/datasets/Chainticks/defi-liquidations): normalized public liquidation events.
- [MEV Tape](https://huggingface.co/datasets/Chainticks/mev-tape): public relay MEV payload metadata.
- [CFTC COT](https://huggingface.co/datasets/Chainticks/cftc-cot): normalized public-domain CFTC Commitments of Traders rows.
- [Funding Divergence](https://huggingface.co/datasets/Chainticks/funding-divergence): derived funding-rate spread examples from publishable inputs.

## Usage Pattern

```python
import pandas as pd

date = "YYYY-MM-DD"
url = "https://huggingface.co/datasets/Chainticks/perp-data/resolve/main/hyperliquid_chain/funding/date={date}/part-0000.parquet"
df = pd.read_parquet(url.format(date=date))
print(df.head())
```

Chainticks is independent and is not affiliated with the protocols, relays, venues, or agencies represented by the data.