Challenge workflow
Start of the challenge
- Create a repository for the code using cookiecutter
- Add dates to the calendar
- Download rules of the challenge
- Bookmark challenge folder on file explorer
- Create a Google keep label for tasks and ideas of the challenge
- Download the challenge data
- Create a conda environment for the challenge and add it to jupyter
conda create -n arc pytest rope pylint tqdm numpy pandas scikit-learn ipython ipykernel coverage ipywidgets matplotlib python=3.10 -y
conda activate arc
python -m ipykernel install --user --name $CONDA_DEFAULT_ENV --display-name "Python ($CONDA_DEFAULT_ENV)"
make env-export
- Create a github repo to have a backup of the data. Vscode allows to do it directly without having to go to the website, choose a private repo. At the end of the challenge it will be made public.
- Use TDD methodology whenever possible, this will save time because errors
won't be propagated along the challenge.
- Have an apprentice attitude, collaborate on the forum, I have a lot to learn
from Kaggle.
- Add a nice picture to README
End of the challenge
- Prepare a report with a summary of the approach to the challenge
- Download the Google keep tasks to the repository in pdf format
- Delete the tasks on google keep and the label
- Delete unnecessary data
- Update the environment yml