This project serves as a template for kickstarting your next data science project. It leverages Cookiecutter to provide a standardized and organized structure for your project, making it easy to manage.
To use the template, just run the following commands:
$ pip install cookiecutter
$ cookiecutter https://github.com/drivendata/cookiecutter-data-science.git├── config -> Configurations for controlling your experiments
├── data -> Data Sources (Don't Push them!)
├── docs -> Documentation of your project
├── figures -> Generated Graphics and Reports
├── models -> Trained Models or Model Cards
├── notebooks -> Notebooks (for initial experimentation)
├── src
│ ├── data -> Scripts for generating or extracting data
│ ├── preprocessing -> Preprocess Data for the models
│ ├── training -> Training scripts to train and evaluate the models
│ └── visualization -> Scripts for visualizing predictions, performances, etc.
└── tests -> Test suite
Here, you will find the most important tools for managing a Python Project + keeping it clean.
- Poetry for Package Management
- Pytest + Useful Plugins for Testing (e.g. hypothesis, pytest-xdist)
- Pre-Commit + Useful Hooks for automating code quality checks and fixes
- Nox for testing your code on multiple Python versions