The aim of this project is to forecast the French electric mix for the next few days.
The electric mix is the relative proportion of each energy source used to produce electricity. The main energy sources are: nuclear, wind, solar, hydraulic, coal, gas, bioenergy and waste.
The Wind and solar energy sources are green, but they are intermittent and depend on the weather. The hydraulic energy source is also dependent on the weather, but to a lesser extent.
Hence, the weather directly impacts the CO2 emissions of the electry consumption. Knowing the CO2 emission forecast can help plan the time of use of electric devices to reduce the carbon footprint.
The project is divided into 2 parts:
We gathered the data from the French electricity transmission system operator RTE (Réseau de Transport d'Electricité) and the French meteorological service Météo France. The data start in 2022 and end in 2024.
We used the data to:
We will use the model trained in the first part to forecast the electric mix for the next few days.
The forecast is updated every day.
A website will be created to display the forecast.
We use hatch to manage the project.
There is a bug in the current version of eccodes, failing the installation.
Installing with conda first solved the issue for me.
Run the following commands to check if eccodes is installed:
python -m eccodes selfcheckYou can use Hatch to activate the virtual environment and install the dependencies with:
hatch shellThis will open a shell with the virtual environment activated.
The dependencies listed in the pyproject.toml file will be installed automatically.
You can run the tests with:
hatch testThis will run the tests in the virtual environment.
You can serve the dashboard with:
hatch serve:prodThis will serve the dashboard on http://localhost:8000.
The dashoard is developed with Taipy. Lauchind the app will automaticaly