Getting Python onto the machine you actually have

There are three starting points, and all of them are free

Most tutorials assume a laptop with an administrator password. Plenty of people reading this have a phone, or a shared computer they cannot install software on. All three cases work.

You already have a laptop or desktop. Install from python.org/downloads. On Windows, the installer shows a checkbox at the bottom of the first screen: Add python.exe to PATH. Tick it. If you do not, the installation succeeds and then every command in every tutorial fails with 'python' is not recognized, and the fix is to run the installer again and choose Modify. This one checkbox is the single most common reason a beginner gives up in the first hour.

On macOS, the installer gives you a command called python3. On Linux, Python is already there; python3 --version will tell you.

You have a phone. On Android, install Pydroid 3 from the Play Store — it ships a real Python with numpy and pandas already compiled, which matters, because compiling them yourself on a phone takes an hour. Termux (from F-Droid, not the Play Store version, which is abandoned) gives you a proper Linux shell and pkg install python. On iPhone, a-Shell is free and includes Python 3.

You have a browser and nothing else. Go to colab.research.google.com and start a new notebook. You get a free machine with Python, numpy, pandas and most of what this course needs, already installed. Kaggle Notebooks are the same idea. The honest limitation: the machine is temporary. It disconnects after roughly 90 minutes of inactivity and at most 12 hours in a session, and when it goes, every file you created on it goes too. Download anything you want to keep, or connect it to Google Drive.

Check that it worked

Open a terminal — Command Prompt or PowerShell on Windows, Terminal on macOS and Linux, the app itself on a phone — and type:

python3 --version

You want something like Python 3.12.4. On Windows the command is usually just python.

The python versus python3 split

This confuses everybody once. Python 2 and Python 3 are different languages with the same name. Python 2 was retired on 1 January 2020, but for twenty years python meant Python 2 on Unix systems, so macOS and most Linux distributions still refuse to let python mean Python 3 by default. Some now leave python pointing at nothing at all.

The rule that always works: type python3 on macOS and Linux, python on Windows. If a tutorial says python and you get an error or a version starting with 2, add the 3.

The same split hits the installer: pip and pip3. Safer than both is python3 -m pip, which installs into the same Python you are running, and cannot get out of step. Lesson nine goes into why that matters.

Which version to install

Take the current stable release minus one. As of writing that means 3.12 rather than the newest. The reason is concrete: large libraries publish precompiled wheels for each Python version, and for a few months after a new release those wheels do not exist yet. pip install pandas then tries to build from source, which needs a C compiler you probably do not have, and fails with 300 lines of red text. Being one version behind avoids a whole class of pain.

Anything from 3.9 upwards runs every example in this course.

An editor, and why not a word processor

You need something that writes plain text. Never Word, Pages or Google Docs — they insert curly quotation marks, and Python does not accept them.

• IDLE ships with Python. It is plain, and it is already installed.
• Thonny (thonny.org) is built for beginners: it shows variables changing as the program runs, which is genuinely useful for your first fortnight.
• VS Code is free, is what most jobs use, and is heavier. Install the Python extension after it.
• On a phone, Pydroid has its own editor.

Any of these is fine. Do not spend an evening choosing.

Make a folder now

Create one folder for this course — Documents/python will do. Put every file you write inside it. When lesson nine arrives with virtual environments, and lesson thirty-something with imports failing, half of those problems are really "which folder was I in". Starting tidy costs nothing.

The setup is not the skill. If installation fights you for more than half an hour, open Colab in a browser and start learning. You can come back and install later, with more context and less frustration.