Credit: Lake Eland Game Reserve

Anaconda is bloated — Set up a lean, robust data science environment with Miniconda and Conda-Forge

In this tutorial, I will describe a process for setting up a lean and robust Python data science environment on your system. By the end of the tutorial, your system will be set up such that:

• Python is installed with only the most common and useful packages for data science
• Conda is installed to manage packages and environments
• You’ll have a single, robust environment which minimizes dependency issues by relying on the conda-forge channel

This tutorial will cover installation for macOS, Linux, and Windows. A video tutorial is also available.

Common Data Science Setup

Currently, the standard advice for someone who desires to get their system set up to data science in Python is to download and install the Anaconda distribution provided by a company with the same name. Installing this distribution with the default settings should allow you to quickly get started doing data…