EnergyAPI: An End-to-End API for Energy-Aware Forecasting and Scheduling

Electricity grids today serve a far more complex mix of consumers than they were designed for. Data centers, EV charging networks, and industrial facilities participate in demand-response programs, shifting flexible workloads to windows where electricity is cheaper or cleaner. Acting on those signals requires software that collects real-time grid data across multiple regions, forecasts how conditions will evolve over the next several hours, and exposes both through a clean API that developers can query directly. Existing open-source infrastructure for grid signal forecasting covers carbon intensity reasonably well, but critical gaps remain: the data acquisition layer and the API serving layer are not connected, fresh data does not automatically trigger model inference, and signals beyond carbon intensity such as demand, price, generation mix, and grid saturation are not exposed through any unified interface. There is also no interactive frontend that makes regional grid conditions spatially visible. EnergyAPI addresses each of these gaps. The platform is structured across five layers: an automated scheduled data acquisition service pulling from NOAA NOMADS, EIA, and ENTSO-E, an extended relational data model covering all five grid signals, a forecasting pipeline connecting fresh data to existing carbon intensity models while adding SARIMA-based demand forecasting and gradient boosting for price, a Django REST API with a workload scheduling interface that accepts SLOs and returns execution plans, and an interactive web frontend with time-series charts and a choropleth map of EU and US regions.