Weather data records and predicts atmospheric conditions such as temperature, precipitation, wind, humidity, pressure and solar radiation. Retailers use it to explain and forecast demand, energy traders to predict load and renewable output, insurers to validate claims, and logistics and agriculture teams to plan operations.
Typical fields
- Location (station ID or latitude and longitude) and timestamp in UTC
- Temperature, dew point and humidity
- Precipitation amount and type
- Wind speed, gusts and direction
- Cloud cover and solar irradiance
- Forecast issue time and lead time for forecast products
Sources and methods
Station observations come from national meteorological services and private networks. They are precise but sparse, and they occasionally have gaps. Gridded products interpolate observations or come from numerical weather prediction models such as those run by NOAA and ECMWF. Reanalysis datasets, such as the Copernicus ERA5 series, combine historical observations with a fixed model to give a consistent gridded history going back decades. Commercial providers often blend several models and apply statistical post-processing to improve local accuracy.
Delivery and pricing
Forecasts and current conditions are usually sold as API plans tiered by calls per month, locations or variables. Historical archives are sold per location, per region or as bulk files in NetCDF, GRIB, Parquet or CSV. Price rises with resolution, update frequency and forecast horizon.
Before you buy
- Ask whether each value is observed, modelled or reanalysis, and how gaps are filled.
- Request forecast verification scores for your locations and lead times.
- Check the time zone and how daily values are aggregated.
- Confirm forecasts are archived as issued if you need to backtest.
- Review licence limits on caching and redistribution.