Load Forecasting Experiments

This project is used to develop simple load forecasting models and apps.

• For an introduction to load forecasting see Introduction to Load Forecasting
• For an introduction to this repo see the 5min video introduction here.

If you just want to play around with the nodebooks and code you can do so via Binder.

• Panel App: https://mybinder.org/v2/gh/MarcSkovMadsen/load-forecasts/HEAD?urlpath=/panel/app (Does not work currently)

You should start from the notebooks/overview.ipynb notebook.

The ERCOT data is proprietary and you would need access to the ERCOT data in order to ingest this data.

Installation

Using conda

conda env create -f environment.yml
conda activate load-forecasts

Notebooks

You can now run

jupyter lab

and open the notebooks/overview.ipynb notebook.

App

You can run the app via

panel serve notebooks/app.py --show

Development

Before you git push please clean the notebooks via

jupyter nbconvert --clear-output --inplace notebooks/*.ipynb

For app development with hot reload use

panel serve notebooks/app.py --auto --show