environments

Starter conda environment files for easy and lightweight data science environment setup

If reproducibility is a big concern, we recommend using instead Docker with our docker images as a starting point.

Available environments

name	description
base-data-science	Basic data science environment, with Jupyter Lab, scikit-learn, pandas, seaborn and standard software engineering utilities
advanced-machine-learning	Same a base-data-science, with the addition of more models, automation and diagnostic tools:xgboost, optuna, yellowbrick and snakemake

Requirements

• conda 3 We recommend using miniconda 3 instead of the full Anaconda 3 distribution

Use

To get started, clone this repository, or download the file for the desired environment

Installing a single environment

Go to the folder of the desired environment file, run

conda env create -f environment.yml

Activating an installed environment

conda activate ENVIRONMENT_NAME

View available environments with

conda env list

Launch Jupyter Lab in that environment with

jupyter lab

Using one of these envrionments as a basis for another

Install the desired environment. Then,

conda create --name NEW_ENVIRONMENT --clone OLD_ENVIRONMENT

License

This repository is licensed under the terms of the MIT License.