Meet our OpenWeather Proprietary Weather model
Resolutionfrom 500 m to 2 km
Data updatesEvery 10 minutes
About OpenWeather Proprietary Weather model
OpenWeather’s advanced ML-based numerical weather model performs at up to 500 m resolution and allows its users to consider highly local nuances of climate and build accurate, efficient products.
The model’s resolution varies from 500 meters for UK territories to 2 kilometers for the other world. The model uses a number of data sources - radars, models from global meteorological agencies (e.g., Met Office, NOAA, ECMWF), weather satellites, and a vast network of weather stations.
Our model is being updated every 10 minutes and provides data on all essential weather parameters such as temperature, precipitation, wind, pressure and others.
Multi-Dimensional Data Storage
We store decades of weather forecasts and measurements history all around the world in a unique multi-dimensional raster data storage solution, which we develop internally. It supports unlimited storage cluster horizontal scalability combined with instant access to any location, time or weather parameter requested. Also this solution is not specific to meteorology domain and operates with user-defined arrays of numeric data, so could be used in other industries.
Real-Time Data Processing
In addition to serving our pre-calculated forecast and historical data we implement on-demand real-time data processing that enables a more precise and rich set of weather parameters to be returned for particular geographical locations.
High Availability & Scalability
Our main goal is to deliver weather data to our customers in the most reliable way exactly at the moment they need it, so we developed uniquely flexible architecture for our API servers, multi-level intermediate caches and backend data storage which allows both to sustain tens of thousands requests from all over the world and rapidly scale out to handle peak loads which may be 2-3 times higher than average.