GridOps - Brazil

The main goal of the app is to share real time information about Brazil's energy matrix and its behaviour during the day.

• Data is requested from ONS (Operador do Sistema Nacional) endpoint.
• Then, data is stored inside a MongoDB time series schema, from the different energy sources and divided by region.

The project contains a machine learning server (ML_server), which contains a Python API endpoint that runs a LSTM forecasting models:

• Apache Airflow tool that re-trains the model on a daily basis, and also generates forecasts.
• Endpoint running on FastAPI to get predictions from the database.

Next steps, gathering enough data to have reasonable predictions from the ML model, and then build the frontend component to render forecasting.