{"id":700,"date":"2022-12-02T10:04:30","date_gmt":"2022-12-02T10:04:30","guid":{"rendered":"https:\/\/tbekk.com\/devstream\/?p=700"},"modified":"2022-12-02T10:06:29","modified_gmt":"2022-12-02T10:06:29","slug":"configuring-your-windows-development-machine-for-data-science","status":"publish","type":"post","link":"https:\/\/tbekk.com\/devstream\/2022\/12\/02\/configuring-your-windows-development-machine-for-data-science\/","title":{"rendered":"Configuring your Windows development machine for data science"},"content":{"rendered":"\n

Interested in topics such as Data Science, Machine Learning, Data Analysis, etc.? Working on windows? Here’s an article explaininng how you can setup your box for such tasks.<\/em><\/p>\n\n\n\n

\n\n\n\n

Introduction<\/h2>\n\n\n\n

In this post, I\u2019ll show you how you can set up your Windows development machine for data science projects. I assume that you have Windows 11 installed on your machine, and no other applications or packages related to data science. I like to configure my machine using VS Code, WSL, Ubuntu, Zsh, Miniconda, and Git, so that\u2019s what I\u2019ll be showing you in this post. If you\u2019re new to data science, I recommend that you follow the same technology choices \u2014 with time, you\u2019ll develop your own preferences. If you\u2019re a seasoned data scientist, you can follow the sections that match your preferences and complement them with your own favorite technologies.<\/p>\n\n\n\n

Let\u2019s get started!<\/p>\n\n\n\n

VS Code<\/h2>\n\n\n\n

VS Code is my editor of choice. It\u2019s great for so many reasons, but especially because it integrates well with the Windows Subsystem for Linux (WSL for short). You can download and install the Windows version of VS Code from this page<\/a>. Once you\u2019ve got VS Code installed and running, choose Extensions in the left navigation pane, search for the \u201cWSL\u201d extension from Microsoft and install it:<\/p>\n\n\n\n

\"Screenshot<\/figure>\n\n\n\n

Windows Subsystem for Linux<\/h2>\n\n\n\n

The next step in our setup is to install WSL<\/a>, which enables us to use Linux directly on a Windows machine, without the overhead of a traditional virtual machine or dual-boot setup. As a former Linux user, WSL has been a game-changer for me on Windows. I\u2019ve been using WSL since its early beta days, and I love how much easier it is to install now, and how stable it is.<\/p>\n\n\n\n

In order to get WSL set up on your machine, you need to enable the \u201cVirtual Machine Platform\u201d feature on Windows. You can do that by typing \u201cTurn Windows features on or off\u201d in the taskbar\u2019s search field:<\/p>\n\n\n\n

\"Screenshot<\/figure>\n\n\n\n

Then check the checkbox near \u201cVirtual Machine Platform\u201d:<\/p>\n\n\n\n

\"Screenshot<\/figure>\n\n\n\n

Next, open the Microsoft Store app, search for \u201cWindows Subsystem for Linux\u201d by \u201cMicrosoft Corporation,\u201d and install it:<\/p>\n\n\n\n

\"Screenshot<\/figure>\n\n\n\n

If you run into issues or need more detailed steps, I recommend that you read this tutorial<\/a>.<\/p>\n\n\n\n

Ubuntu<\/h2>\n\n\n\n

You can also use the Microsoft Store to install whichever Linux distribution you want to use with WSL. I generally use Ubuntu, so that\u2019s what I\u2019ll show here. If you search for \u201cUbuntu\u201d in the Microsoft store, you might see something similar to this:<\/p>\n\n\n\n

\"Screenshot<\/figure>\n\n\n\n

You can pick a specific version of Ubuntu, but I prefer to choose the option without a version, which automatically upgrades to the latest stable version.<\/p>\n\n\n\n

Once Ubuntu finishes installing, open the app. This launches a terminal window, where you\u2019ll be prompted to supply a username and password for the newly created Linux administrator account. These are the credentials you\u2019ll need to enter whenever you execute a command with \u201csudo\u201d to elevate privileges, so choose wisely.<\/p>\n\n\n\n

Zsh<\/h2>\n\n\n\n

The Zsh<\/a> shell is an alternative to the default Bash<\/a> shell. If you\u2019re happy with Bash and want to keep your machine setup simple, you can skip this section. I like to use Zsh because it supports everything that Bash supports, with several useful additional features: it has much smarter support for tab auto-completion, it includes many commonly used aliases, and it shares command history across active shells.<\/p>\n\n\n\n

If you don\u2019t already have an Ubuntu terminal window open, search for and start the \u201cTerminal\u201d app. Click on the downward-pointing chevron in the title bar and select the \u201cUbuntu\u201d option to open a new tab that\u2019s running Ubuntu. To install Zsh, type the following command and enter your administrator password if you\u2019re asked for credentials:<\/p>\n\n\n\n

sudo apt-get install zsh\n<\/code><\/pre>\n\n\n\n
The first time you run Zsh, you will be asked how you want to create your .zshrc<\/code> configuration file:<\/p>\n\n\n\n
zsh\n<\/code><\/pre>\n\n\n\n
\"Screenshot<\/figure>\n\n\n\n
I generally select option 0<\/code>, which creates the .zshrc<\/code> file with just a comment. We\u2019ll customize this file in a later step.<\/p>\n\n\n\n
Next, I like to install Oh My Zsh<\/a>, which is an open source, community-driven framework for managing your Zsh configuration. It provides you with a wide range of plugins and themes, and it has a vibrant community of users. It\u2019s my framework of choice for managing Zsh. Here\u2019s how we can install it:<\/p>\n\n\n\n
sh -c \"$(curl -fsSL https:\/\/raw.github.com\/ohmyzsh\/ohmyzsh\/master\/tools\/install.sh)\"\n<\/code><\/pre>\n\n\n\n
We\u2019ll configure Oh My Zsh later in this post, by adding a few plugins to the .zshrc<\/code> configuration file.<\/p>\n\n\n\n
Miniconda<\/h2>\n\n\n\n
When coding with Python, it\u2019s convenient to have a tool that manages our packages and environments for each project. It\u2019s common for different projects to use different versions of the same packages, which can make switching from one project to the next unwieldy. That\u2019s where tools like Miniconda<\/a> come in: they enable us to switch from one set of dependencies to another with a single command. There are many other package managers with similar goals, such as Conda<\/a>, pip<\/a>, and venv<\/a>. I prefer Miniconda to Conda because I don\u2019t need most of the extra functionality that comes with Conda, such as support for programming languages other than Python. And I like both Miniconda and Conda better than other package managers I\u2019ve used because they help me manage the version of Python itself, in addition to all other packages that my project needs.<\/p>\n\n\n\n
You can find the installation links for Miniconda on this page<\/a>. For a WSL setup, you\u2019ll need to choose the \u201cMiniconda3 Linux 64-bit\u201d version. Clicking on the link will download a file with extension .sh<\/code>. You can install it by copying it to a suitable location within Ubuntu, and then running the following command:<\/p>\n\n\n\n
sh .\/Miniconda3-latest-Linux-x86_64.sh\n<\/code><\/pre>\n\n\n\n
You need the Linux version, rather than the Windows version, because you\u2019ll be doing your package management from within WSL, not Windows!<\/p>\n\n\n\n
Git<\/h2>\n\n\n\n
I use Git (in conjunction with GitHub) for source code management \u2014 who doesn\u2019t, these days? I like to be able to run Git commands both from within VS Code (which runs on Windows) and from the terminal (which is running Linux). I prefer VS Code\u2019s user interface for common operations like creating branches, committing and pushing changes, and pulling the latest code; but I still want access to Git in my Linux terminal for more advanced operations. Therefore, I configure Git twice \u2014 once for Windows and once for Linux.<\/p>\n\n\n\n
Thankfully the steps you need to follow on Windows and Linux are pretty much the same, the only difference being that on Windows you execute them using the Git Bash app, and on Linux you execute them within your terminal. The GitHub web site provides excellent documentation on how to set up Git, so I\u2019ll just link to the appropriate pages. Note that these pages contain steps for both Windows and Linux, so be sure to follow the instructions on each of those tabs:<\/p>\n\n\n\n