Python Tutorial: Anaconda - Installation and Using Conda

Anaconda is pitched as a practical shortcut for getting a working Python data-science setup—complete with common libraries, environment management, and a package manager—without manually stitching everything together. Instead of feeling forced to abandon pip, the core value is convenience: many beginners run into installation and dependency problems, and Anaconda bundles a large, data-focused toolkit (including numpy and Jupyter Notebook) plus conda for managing packages and environments.

The tutorial starts by clarifying what Anaconda actually includes: a Python distribution, the conda package manager, tools for separating environments, and a broad set of pre-installed libraries geared toward data science. It then addresses the most common question—why use Anaconda at all. The answer is pragmatic rather than ideological: if someone is already comfortable installing packages with regular Python and pip, there’s no urgent need to switch. But for data science learners, classrooms, and teams that need consistent setups across machines, Anaconda reduces friction around installing Python on multiple platforms, handling permissions, and ensuring the right package versions are available.

Installation begins at the Anaconda download page, where users can check the list of pre-installed packages. The tutorial notes that the full Anaconda distribution can take significant disk space because it includes many libraries up front. For those with storage constraints, a smaller alternative called “mini conda” is mentioned as a bootstrap option that installs essential components (Python distribution essentials and conda), though the walkthrough proceeds with the full installer.

After downloading the macOS graphical installer, the process is largely “click through defaults.” Once installed, the tutorial verifies success by checking that Python is on the system path and that the active Python comes from the Anaconda distribution. It confirms package availability by importing numpy and matplotlib in a Python session. From there, it shows how pip still works for adding packages later, using pip list to inspect what’s already installed.

The focus then shifts to conda as Anaconda’s built-in package manager. Commands like conda list, conda search, and conda install are presented as close cousins to pip’s workflow, with conda also capable of installing non-Python dependencies. The tutorial highlights conda’s environment management as a major differentiator: separate projects can require different versions of Python and libraries, so isolated environments prevent version conflicts.

To demonstrate, it creates a new environment named “my app” with starting packages flask and sql alchemy, then activates it (using source activate on macOS/Linux). Inside that environment, pip list shows only the environment’s packages, and which Python confirms the interpreter path is now environment-specific. It also covers deactivation (source deactivate). A second environment, “my app 2.7,” is created using python=2.7 to show how conda can pin a different Python version for the same project stack.

Finally, the tutorial shows how to list environments with conda env list and remove them with conda remove, then ties the setup back to real use cases: Jupyter Notebook is commonly recommended alongside Anaconda because, after installation, the command jupyter notebook can launch notebooks immediately. The takeaway is that Anaconda isn’t mandatory, but it can be a strong starting point for data science workflows where consistency and dependency management matter.