Conda: Package & Environment Manager

Overview Conda is an open-source package management system and environment management system. Unlike pip (which only manages Python), Conda can install any software (C libraries, R packages, Compilers, GPU drivers).

Anaconda vs Miniconda

• Anaconda: A massive distribution (3GB+) containing Python + 1,500 scientific packages (Jupyter, Pandas, NumPy) pre-installed. Good for beginners.
• Miniconda: A minimal installer (50MB) containing only Conda and Python. You install what you need. Good for pros/servers.

Common Commands

# Create environment with specific python version
conda create --name myenv python=3.9

# Activate
conda activate myenv

# Install (from Anaconda channel)
conda install numpy

# Install (from conda-forge, the community channel)
conda install -c conda-forge opencv

Conda vs Pip

• Pip: Compiles form source often. Fails if you lack C compilers.
• Conda: Installs pre-compiled binaries. "It just works" regarding complex dependencies like CUDA or GDAL.

Data Scientists prefer Conda; Web Developers prefer Pip.