Mastering Conda Environments

If you’re working with Python for data science or machine learning, you’ve likely encountered Conda. It’s not just another package manager—it’s a powerful environment manager that solves the notorious “it works on my machine” problem. Today, I’m sharing my workflow for keeping multiple projects neatly isolated using Conda environments.

Getting Started with Conda

If you haven’t installed Conda yet, I recommend the minimal Miniconda distribution. Once installed, verify it’s working:

conda --version

Creating Custom Environments

The beauty of Conda is creating separate environments for different projects. Here’s how I typically set up a new environment:

conda create --name myenv python=3.8

This creates an environment named “myenv” with Python 3.8. I usually specify the Python version to avoid surprises later.

Activating and Deactivating Environments

Before installing packages or running code, activate your environment:

conda activate myenv

When you’re done, you can deactivate it:

conda deactivate

Managing Packages

Once your environment is active, you can install packages using either conda or pip:

conda install numpy pandas matplotlib
# or
pip install tensorflow keras

I generally prefer using conda for packages available in Conda repositories, falling back to pip only when necessary.

Viewing Your Environments

To see all environments you’ve created:

conda env list

To check what packages are installed in your current environment:

conda list

Exporting and Recreating Environments

This is where Conda really shines for collaboration. You can export your environment configuration:

conda env export > environment.yml

Then anyone can recreate the exact same environment:

conda env create -f environment.yml

I always include the environment.yml file in my project repositories to make it easier for collaborators (and my future self).

Removing Environments

Spring cleaning for your development setup:

conda env remove --name myenv

Advanced Tips

  • Creating environments in specific locations: Use the --prefix option to create an environment in a project directory: 
    conda create --prefix ./env python=3.8
  • Updating all packages: Keep your environment up-to-date: 
    conda update --all
  • Clean up unused packages: Free up disk space: 
    conda clean --all

Proper environment management has saved me countless hours of debugging dependency conflicts. What’s your Conda workflow? I’d love to hear your tips and tricks!