cookiecutter-datascience-simple

Cookiecutter template for "simple" datascience projects. These projects are characterized by the following:

• primarily for single developer
• its conda virtual environment is based on some existing activated environment
• most analysis done in notebooks
• may be some Python code in .py files
• want easy importing of user created Python code into notebooks without install
• no intention of creating distributable package
• no automated testing (e.g. pytest)
• simple folder structure
• basic documentation is initialized with a markdown file

This cookiecutter was developed for use in teaching a Python based analytics course that includes some basic software engineering content. One of the first topics we talk about is project structure and I wanted a very simple cookiecutter to use for quickly setting up a reasonable project folder structure.

More complex cookiecutters will be developed and used for use in the course as well when talking about, say, creating deployable Python packages.

The cookiecutter package is already installed in the conda virtual environment we are using in class. See the installation instructions if you need to install it in some other virtual environment.

Usage

To start a new project from this cookiecutter, first activate your desired conda environment. Then, do the following from the directory into which a new project directory is to be created:

cookiecutter gh:misken/cookiecutter-datascience-simple

Then what?

Navigate into your newly created project directory and read the README in there. Do those things and then get to work.