Skip to content

Challenge workflow

Start of the challenge

  1. Create a repository for the code using cookiecutter
  2. Add dates to the calendar
  3. Download rules of the challenge
  4. Bookmark challenge folder on file explorer
  5. Create a Google keep label for tasks and ideas of the challenge
  6. Download the challenge data
  7. 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
  1. 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.
  2. Use TDD methodology whenever possible, this will save time because errors won't be propagated along the challenge.
  3. Have an apprentice attitude, collaborate on the forum, I have a lot to learn from Kaggle.
  4. Add a nice picture to README

End of the challenge

  1. Prepare a report with a summary of the approach to the challenge
  2. Download the Google keep tasks to the repository in pdf format
  3. Delete the tasks on google keep and the label
  4. Delete unnecessary data
  5. Update the environment yml