Overview

PMData historical data is delivered as Apache Parquet files. Schema pages define the fields and row semantics of each downloaded file.

Available schemas

Schema	Data type	Description	Use case
L2	l2	Incremental L2 order book and market events	Order book replay and liquidity analysis
Trades	trades	Polymarket WebSocket trade executions	Price, volume, and trade-flow analysis
On-Chain Fills	onchain_fills	Polygon fills with maker and taker wallets	Wallet and on-chain flow analysis
Chainlink Streams	streams	Raw price data received from the Chainlink WebSocket stream, not the Polymarket proxy.	Reference-price and latency analysis
Chainlink TWAP	streams_twap30s, streams_twap60s	TWAP 30s/60s data received from the Chainlink WebSocket stream.	Resolution and benchmark analysis

Common fields

Polymarket WebSocket files in the l2 and trades schemas share these fields:

Field	Type	Description
market_slug	string	Polymarket market slug.
timestamp	timestamp[us]	Polymarket event timestamp.
local_timestamp	timestamp[us]	PMData receive timestamp.
event_type	string	Source event represented by the row.

On-Chain Fills and Chainlink use source-specific fields described on their schema pages.

Timestamps

All timestamps use UTC and microsecond precision.

• timestamp in L2 and Trades is the timestamp supplied by Polymarket.
• local_timestamp is when the PMData collector received the WebSocket event.
• timestamp in On-Chain Fills is the Polygon block timestamp.
• observationsTimestamp in Chainlink is the source observation time.
• receiveMicrosecondTimestamp is when PMData received the Chainlink report.

The difference between a source timestamp and its PMData receive timestamp represents delivery and processing latency.

File conventions

• L2 contains several event types in one schema. Fields that do not apply to a row's event_type are null.
• Order book prices and sizes use Parquet list columns.
• Chainlink price values are stored as scaled integers in string columns. See the Chainlink schema.