Forecasting Workflow

Project for forecasting / prediction tasks.

Purpose of this project

• To get familiar with other types of data (the author has been focusing on computer vision problems only).
• To understand better how to implement the following libraries:
  • Pandas
  • SQL (sqlalchemy, mysql-connector-python)
  • Traditional ML (scikit-learn, xgboost)
  • Training transformers using PyTorch
• Practice implementing tests better.
• Explore UV as the new package (and project/repo) manager.
• Try tools like mypy, ruff, etc. for better development experience.

Takeaways

• UV is a lot better than PIP. Will migrate to UV in the future.
• Will start to employ more tools for cleaner codes, better coding experience, and more efficient debugging process.

Others

• mkdocs implementation can be found here.