Case Study: Hive Power achieves superior energy forecast accuracy and improved ML performance with Meteomatics Weather API

Higher Accuracy Leads to Improved Machine Learning for Energy Forecasting! Hive Power Reveals the Results of Its Weather Forecast Verification

Hive Power, a Swiss smart‑grid SaaS provider, needed highly accurate short‑term weather forecasts to train and run its machine‑learning energy prediction models (typical horizons 2–48 hours). After benchmarking a dozen providers using a year of hourly temperature and solar radiation forecasts versus observations, Hive Power selected Meteomatics and its Meteomatics Weather API for its superior accuracy and the availability of archived forecast data (back to 1978), which is critical for robust model training.

Meteomatics’ Weather API was integrated into Hive Power’s Forecaster, delivering the most accurate, well‑calibrated temperature and solar radiation forecasts (narrow, zero‑centred error distributions and lower MAE compared to competitors). That higher weather accuracy translated into better energy generation and consumption forecasts, enabling optimized asset orchestration, tangible cost savings, reduced planning risk, and improved peak‑shaving, trading performance and grid stability for Hive Power.

Hive Power

Gianluca Corbellini

Managing Director