A package to create electricity demand prediction for any arbitrary shape using ML
Training data:
- histortic and future hourly temperature time series
- histortic and future hourly humidity time series
- hourly wind speed
- yearly GDP
- ...
General note:
- Data should be available by default globally, in high resolution rasters (.nc format), over 30+year
- Methods should be written that more accurate local data can be integrated
- Methods should allow to modify flexibly climate change impacts on the demand
Possible future? Create demand timeseries for other sectors such as heat demand, industrial demand and transportation demand
Snakemake exercise
- Install pypsa-africa or pypsa-eur as conda environment, see here
- Run the following to get an output
conda activate pypsa-africa
snakemake --cores 1 salt_bae_extractor
- Try to understand how this output was generated in the Snakefile
- Analyse the python scripts
- Play around with the config.yaml