Python installation depends on your OS (operating system). As there's quite a bit of shell language involved in the set-up, here I also introduce two alternatives:
- A user-friendly Python and R distribution plus package manager, Anaconda.
- A cloud-based alternative, Google Colab, which should be sufficient for the purposes of this course. That being said, configuring your own Python environment is still encouraged.
Sections:
In the Windows start menu search bar, type Python to see which versions you have installed. If it's already 3.11, skip to section Python packages via PIP; if not, read on to install Python 3.11
Go the the official Python website. Select "Windows installer 64-bit" (if an error occurs and the installation fails, your devic is on a 32-bit processor, so switch to "Windows installer 32-bit"). After the installer file has been downloaded, double-click on the file to run it. In the pop out window, check off ☑️ "Install launcher for all users (recommended)" and ☑️ "Add Python 3.11 to PATH", then click "Install Now".
Launch your PowerShell, type
py --list
to show the Python versions that you have installed (and have been added to PATH). You should at least see 3.11.
- 💡 To launch PowerShell, type "powershell" in the Windows start menu search bar and click on the PowerShell icon; if you're on Windows 11, press the Windows key and X.
In the PowerShell, type
py -3.11 -m pip list
to see all the packages installed by PIP. If your Python 3.11 has just been installed, you'll only see pip itself and setuptools. In the following steps, we'll install a package called virtualenv to create a virtual environment:
py -3.11 -m pip install virtualenv
After installation, create a new folder called erpclass, go to that folder, create a virtual environment named venv, and install relevant packages by executing the commands below one line at a time:
mkdir erpclass; cd erpclass
py -3.11 -m virtualenv venv
venv\Scripts\activate
pip install jupyterlab mne pandas pingouin openpyxl flake8
- MNE is the Python package for reading, processing, analyzing, and plotting EEG/MEG data.
- Pandas is a powerful library for data analysis and data manipulation in the form of
pandas.DataFrame - Pingouin is for statistical analyses in Python (I find it more convenient than Statsmodels, which comes installed as a dependency with MNE.
- Openpyxl is needed in order for Python to read from and write to Excel files.
- Side note: Excel now accepts (or is about to accept) Python syntax and functions 😍
- Flake8 is basically Grammerly for Python.
⚠️ In the third line, if you encounter an error related to execution policies, launch a new PowerShell window and run as the administrator. Type
Set-ExecutionPolicy Unrestricted
then type
cd C:\Users\<your_user_name>\erpclass; venv\Scripts\activate
(change <your_user_name> to your actual device user name)
to see if the error gets resolved. If so, proceed to the fourth line
pip install jupyterlab mne pandas pingouin openpyxl flake8
⚠️ In the fourth line, ifpipdoesn't work, perhaps trypy -m pip.
After it's done, type pip list (or py -m pip list) again. You should now see a long list of packages installed.
No we're ready to open your Jupyter Lab! Type
jupyter-lab
which will automatically open a browser tab (if it doesn't, copy and paste any one of the URLs, e.g., http:/localhost:blahblahblah into your browser search bar), and voila! There's your Jupyter Lab! Note also that you do not need the internet to run Jupyter; localhost refers to your device.
Last step, open a Jupyter notebook (in the Launcher, select "Pytnon 3 ipython kernel") and paste the following lines into the first cell, then hit Run:
import mne
import numpy as np
from platform import python_version
python_version()
There should be no errors, and 3.11.5 will be the output.
To deactivate a virtual environment, simply type
deactivate
Suppose you've closed your terminal and wish to relaunch it, you have to go to the erpclass folder and activate venv, after which you may then launch Jupyter:
cd erpclass
venv/Scripts/activate
jupyter lab
Since most macOS computers come with Python pre-installed (please do not modify your mac's pre-installed Python as it would mess with your system), it's always a good practice to check which Python version you have on your device. Go to your terminal and type python --version which probably isn't 3.11.5. The steps below will install the latest Python release and a virtual environment for better package version/dependency management.
- 💡 To launch the mac terminal, click on the Launchpad icon in the Dock, type
Terminalin the search field, then click Terminal. - 💡 Apple installs Python in
usr/bin/, so don't touch anything related to Python in that directory!!!
Check if they're installed:
xcode-select --version
If they are, you'll get the corresponding version in your terminal output; if not, run
xcode-select --install
which will pop out a window asking you to install them. Click Install, and wait. You will then see a message display
The software was installed.
Click Done, and type
xcode-select -p
in the terminal, which should output
/Library/Developer/CommandLineTools
In the terminal, type
brew --version
to see if Homebrew is installed on your mac. If an error occurs, it is not installed. Type the following command to install Homebrew:
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
After installation, type
brew doctor
to check that everything's okay.
- 💡 For an alternative way to install Python with less terminal code involved, please refer to this page.
After Homebrew has been installed, run the following lines one line at a time in the terminal to install Python 3.11.5 and the packages we'll be using in class, along with several others I find useful for data analysis:
brew update && brew upgrade
brew install pyenv
After installation, run
pyenv init
and look at the terminal output carefully; follow the instruction there, i.e., copy and paste the listed lines to your corresponding files, which in my case are ~/.bashrc and ~/.profile as I'm running Ubuntu in Windows Subsystem for Linux. Remember to save, then source the files or close and restart your terminal for the changes to take effect.
- 💡 If you've never modified such files and are lost at this step, feel free to screenshot what
pyenv initoutputs and email me the picture.
It's recommended to install build depedndencies before installing any Python version, so execute the following:
brew install openssl readline sqlite3 xz zlib tcl-tk
Now install Python 3.11.5 using pyenv (this may take a while):
pyenv install 3.11.5
Set Python to the newly installed version:
pyenv global 3.11.5
Type
which python
which should output something like
blahblahblah/.pyenv/shims/python
And typing
python --version
should get you the desired 3.11.5
Now type
pip3 list
to see the packages installed by PIP, which in this case are probably just pip itself and setuptools. We want to install a package called virtualenv to create a virtual environment as mentioned in the beginning, so run
pip3 install virtualenv
After installation, create a new folder called erpclass, go to that folder, and create a virtual environment named venv by executing the commands below one line at a time:
mkdir erpclass; cd erpclass
virtualenv venv
source venv/bin/activate
pip3 install jupyterlab mne pandas pingouin openpyxl flake8
- MNE is the Python package for reading, processing, analyzing, and plotting EEG/MEG data.
- Pandas is a powerful library for data analysis and data manipulation in the form of
pandas.DataFrame - Pingouin is for statistical analyses in Python (I find it more convenient than Statsmodels, which comes installed as a dependency with MNE.
- Openpyxl is needed in order for Python to read from and write to Excel files.
- Side note: Excel now accepts (or is about to accept) Python syntax and functions 😍
- Flake8 is basically Grammerly for Python.
After it's done, type pip3 list again. You should now see a long list of packages installed.
No we're ready to open your Jupyter Lab! Type
jupyter lab
which will automatically open a browser tab (if it doesn't, copy and paste any one of the URLs, e.g., http:/localhost:blahblahblah into your browser search bar), and voila! There's your Jupyter Lab! Note also that you do not need the internet to run Jupyter; localhost refers to your device.
Last step, open a Jupyter notebook (in the Launcher, select "Python 3 ipython kernel") and paste the following lines into the first cell, then hit Run:
import mne
import numpy as np
from platform import python_version
python_version()
There should be no errors, and 3.11.5 will be the output.
To deactivate a virtual environment, simply type
deactivate
in the terminal. In addition, if you now type
pip3 list
after venv has been deactivated, you won't see any of the packages we just installed after virtualenv.
Suppose you've closed your terminal and wish to relaunch it, you have to go to the erpclass folder and activate venv, after which you may launch Jupyter:
cd erpclass
source venv/bin/activate
jupyter lab
Go to the official Anaconda website and download the Graphical Installer for your OS. This may take a while... Check off 3 of the 4 boxes as shown in the screenshot. Also, make sure that there's no existing Python 3.11 on your device (you'll see a warning pop-up if this is the case).
After installation, launch Anaconda Navigator. Most likely you'll see a prompt suggesting that you update Anaconda to the latest version. Click update, which will then prompt you to quit Anaconda Navigator. Just click Yes/Next/Continue all the way and relaunch Anaconda Navigator when the update finishes.
On the Anaconda Navigator graphical interface, select "Environments", click on "Create", then type in venv and wait for the new virtual environment to be created. On the right side of the page, click on "Channels", add conda-forge. In the drop-down menu, select "Not installed", then search for jupyterlab in the top-right search bar, check it off and "Apply".
Continue searching for then applying the following packages:
- mne is the Python package for reading, processing, analyzing, and plotting EEG/MEG data.
- pandas is a powerful library for data analysis and data manipulation in the form of
pandas.DataFrame - pingouin is for statistical analyses in Python (I find it more convenient than Statsmodels, which comes installed as a dependency with MNE.
- openpyxl is needed in order for Python to read from and write to Excel files.
- Side note: Excel now accepts (or is about to accept) Python syntax and functions 😍
- flake8 is basically Grammerly for Python.
If your Anaconda unfortunately gets forever stuck at package installation, hit "Cancel". Select "Home" on the right, launch JupyterLab, which will open a notebook "Untitled1.ipynb". In the first cell, type
!pip install mne pandas pingouin openpyxl flake8
from IPython.display import clear_output
clear_output()
!pip show mne pandas pingouin openpyxl flake8
In the next cell, type
!mkdir erpclass
to create a folder for this class. Now you're all set :)
Google Colab provides free access to IPython notebooks, which can be saved and shared as can any other Google documents. To open a new Colab notebook, log into your Google account, then visit this link. Click on "NEW PYTHON 3 NOTEBOOK" in blue at the bottom right.
Since the current Python version Colab runs on is 3.10.12, and the package MNE for EEG/ERP data analysis does not come pre-installed, make a copy of this notebook to your Google account to configure your Colab environment.