Historical Polymarket Weather Data

Weather contracts often evolve through a sequence of forecast cycles rather than a single breaking event. As the observation window approaches, forecast uncertainty usually changes and new model runs can alter the information available to traders. Tick-level market data lets you examine probability, spread, depth, and trading activity around those update windows without reducing the market to only an opening and closing price.

Start with the contract's exact location, measurement, threshold, observation period, and resolution source. Markets that look similar can resolve from different stations or definitions, so grouping by a broad label such as temperature or precipitation can hide material differences. Preserve the Polymarket slug as the market identifier, use UTC for joins, and record the forecast issue time rather than only the time a forecast was downloaded.

PMData provides historical Polymarket order books, trades, and on-chain fills; it does not provide meteorological observations or numerical weather prediction output. A calibration study should join the market history to an authoritative forecast and observation source chosen for the contract. Keeping market probabilities separate from weather measurements makes it possible to test forecast reactions, time-to-resolution behavior, and resolved-outcome calibration without confusing the traded signal with the underlying physical data.