Setting Up Your Data Science Tool Stack
Read and complete this guide as part of your first homework. We will also be reviewing this in class.
Your stack:
If you already have these installed, you can skip the setup
- R and RStudio: the R language and its dedicated editor
- Python and Conda: the Python language and its package/environment manager
- VS Code: a general editor for Python, notebooks, and Quarto
- Quarto: a publishing system for reports, slides, and websites
- Jupyter: interactive notebooks that mix code and text
Where to go:
- macOS users: the section “macOS Setup”
- Windows users: choose “Native Windows”, “WSL”, or both
Overview and Core Concepts
What is an interpreter?
- An interpreter is a program that reads your source code and runs it directly, line by line, at run time.
- There is no separate “build” step: you hand it a script and it executes right away.
- Contrast: compiled languages (C, C++) are first translated into a standalone executable, then run.
- R and Python are interpreted languages: your code runs through the
Rorpythoninterpreter program. - Takeaway: you must install the interpreter; that is the thing that actually runs your code.
Common file types: Markdown, .qmd, and .ipynb
- Markdown (.md): plain text with lightweight formatting;
#for headers,**bold**,-for lists; renders to HTML or PDF. - Jupyter notebook (.ipynb): a file of cells that mixes Markdown (text) and code; outputs are saved inline; stored as JSON.
- Quarto document (.qmd): plain-text Markdown plus executable code chunks; rendered by Quarto to HTML, PDF, Word, or slides.
- Shared idea: all three combine human instructions with code, so narrative, analysis, and results live in one document.
- Key difference: .ipynb is cell-based JSON; .qmd is clean Markdown text and works with R, Python, and Julia.
- Scripts:
- .py: a Python script file (
.py), which runs python code line by line. - .R: an R script file (
.R), which runs R code line by line.
- .py: a Python script file (
Environments and Conda
- Conda is both a package manager and an environment manager, used mainly for Python.
- We install it through Miniforge, a minimal Conda that uses the free conda-forge package channel by default.
- An environment is an isolated set of packages and versions, so one project does not break another.
- We will create one course environment, named
dsan, and use it everywhere.
The plan (all platforms)
- Install the language interpreters: R, and Python by way of Conda (Miniforge).
- Install the editors: RStudio and VS Code.
- Install Quarto and the VS Code extensions.
- Create one Conda environment, then verify everything runs.
macOS Setup
Before you start (macOS)
- Most current Macs use Apple Silicon (M1 through M4). Check with the Apple menu, then “About This Mac”.
- Apple Silicon: download the “Apple silicon” or “arm64” installers.
- Older Intel Macs: download the “Intel” or “x86_64” installers.
- Your terminal shell is zsh by default; the installers configure it for you.
Step 1: Install R, then RStudio
Check first if you already have R and RStudio installed: in Terminal run
R --version(for R) andls /Applications | grep -i RStudio(for RStudio). Install whichever is missing.
- Install R before RStudio so RStudio can find it.
- R: go to cran.r-project.org, open Download R for macOS, choose the
.pkgthat matches your chip, then run it. - RStudio: download RStudio Desktop, open the
.dmg, drag RStudio into Applications. - Open RStudio; it should detect R automatically.
Step 2: Install Miniforge (Python)
Check first if you already have Conda installed: in Terminal run
conda --version. If you see a version, skip this step.
- Download the macOS installer from conda-forge.org/download, choosing the
.pkg(graphical) for your chip. Command-line alternative:
curl -L -O https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-arm64.sh
bash Miniforge3-MacOSX-arm64.sh
- On Intel Macs, use the
x86_64file instead. Reopen Terminal, then runconda --version. - conda-forge is already the default channel, so no further channel setup is needed.
Step 3: Install VS Code and extensions
Check first if you already have VS Code installed: run
code --version, or look for Visual Studio Code in your Applications folder. If present, skip the download and go straight to the extensions.
- Download VS Code from code.visualstudio.com/download, open the
.zip, move “Visual Studio Code” into Applications. - Open VS Code, go to the Extensions panel, and install: Python, Jupyter, Quarto. Optional: R.
- Optional convenience: Command Palette (Cmd+Shift+P), run “Shell Command: Install ‘code’ command in PATH”.
Step 4: Install Quarto
Check first if you already have Quarto installed: run
quarto --version. If you see a version, skip this step.
- RStudio already bundles Quarto; for VS Code, install the standalone Quarto CLI as well.
- Download the macOS
.pkgfrom quarto.org/docs/download and run it. - Verify in Terminal:
quarto check
Step 5: Create the course environment
Check first if you already have the
dsanenvironment: runconda env list. Ifdsanis listed, skip creating it and just runconda activate dsan.
- In Terminal:
conda create -n dsan python=3.12 jupyterlab pandas numpy matplotlib
conda activate dsan
- In VS Code, open a notebook or Python file, then select the environment: Command Palette, “Python: Select Interpreter”, choose
dsan.
Verify (macOS)
Run each command; you should see version numbers with no errors:
R --version
python --version
conda --version
quarto --version
Then open RStudio, open VS Code, create a test .ipynb, and run a cell.
Windows Setup (Part 1): Native Windows
Two paths on Windows
- Native Windows: install everything directly on Windows; this is the simplest way to start.
- WSL (Ubuntu Linux inside Windows): a Linux environment that matches real servers; covered in Part 2.
- You do not need both. Start with native Windows unless your instructor asks for WSL.
Step 1: Install R, then RStudio
Check first if you already have R and RStudio installed: in PowerShell run
dir "C:\Program Files\R"(for R) anddir "C:\Program Files\RStudio"(for RStudio). A listed folder means it is installed. Install whichever is missing.
- Install R before RStudio so RStudio can find it.
- R: go to cran.r-project.org, open Download R for Windows, base, then run the
.exeand accept the defaults. - RStudio: download RStudio Desktop, run the
.exeinstaller. - Open RStudio; it should detect R automatically.
Step 2: Install Miniforge (Python)
Check first if you already have Conda installed: open the Start Menu and look for “Miniforge Prompt” (or an existing “Anaconda Prompt”). If one exists, Conda is already installed; skip this step.
- Download the Windows
.exefrom conda-forge.org/download. Direct file: Miniforge3-Windows-x86_64.exe. - Run it, “Just Me”; leave “Add to PATH” unchecked.
- Use “Miniforge Prompt” from the Start Menu whenever you run conda commands, then run
conda --version. - conda-forge is already the default channel, so no further channel setup is needed.
Step 3: Install VS Code and extensions
Check first if you already have VS Code installed: in PowerShell run
code --version. If you see a version, skip the download and go straight to the extensions.
- Download the Windows “User Installer” from code.visualstudio.com/download, then run it (no admin needed).
- In the VS Code Extensions panel, install: Python, Jupyter, Quarto. Optional: R.
Step 4: Install Quarto
Check first if you already have Quarto installed: in the Miniforge Prompt run
quarto --version. If you see a version, skip this step.
- RStudio already bundles Quarto; for VS Code, install the standalone Quarto CLI as well.
- Download the Windows
.msifrom quarto.org/docs/download and run it. - Verify:
quarto check
Step 5: Create the course environment
Check first if you already have the
dsanenvironment: in the Miniforge Prompt runconda env list. Ifdsanis listed, skip creating it and just runconda activate dsan.
- In the Miniforge Prompt:
conda create -n dsan python=3.12 jupyterlab pandas numpy matplotlib
conda activate dsan
- In VS Code: Command Palette (Ctrl+Shift+P), “Python: Select Interpreter”, choose
dsan.
Verify (native Windows)
Run in the Miniforge Prompt:
R --version
python --version
conda --version
quarto --version
Common fix: if conda is “not recognized” in a plain PowerShell window, use the Miniforge Prompt instead, or run conda init powershell once and reopen the window.
Windows Setup (Part 2): WSL (Ubuntu)
What WSL is
- WSL runs a real Ubuntu Linux inside Windows; you get a Linux command line and Linux tools.
- It is a good match for servers and cloud environments, and is widely used in data science.
- You will edit code with VS Code on Windows, connected into Ubuntu.
- Reference: Microsoft’s WSL install guide.
Step 1: Install WSL
Check first if you already have WSL installed: in PowerShell run
wsl -l -v. If Ubuntu is listed, skip to Step 2.
- Open PowerShell as Administrator (right-click, “Run as administrator”).
- Install WSL together with Ubuntu:
wsl --install
- Restart your computer when prompted.
- On the first launch of Ubuntu, create a UNIX username and password (separate from your Windows login).
- If WSL fails to start, enable virtualization in your BIOS or UEFI settings, then retry.
Step 2: Update Ubuntu
- In the Ubuntu terminal, refresh the system packages:
sudo apt update && sudo apt upgrade -y
Step 3: Connect VS Code to WSL
Check first if you already have VS Code installed: in Windows PowerShell run
code --version. If missing, install VS Code on Windows (Part 1, Step 3).
- In VS Code, install the WSL extension.
- From Ubuntu, make a project folder and open it in VS Code:
mkdir -p ~/dsan && cd ~/dsan
code .
- VS Code reopens “connected to WSL”; when prompted, install the Python, Jupyter, and Quarto extensions inside WSL too.
Step 4: Install Miniforge in Ubuntu
Check first if you already have Conda installed: in Ubuntu run
conda --version. If you see a version, skip this step.
- Download and run the Linux installer (use the
aarch64file on Arm-based Windows):
cd ~
wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh
bash Miniforge3-Linux-x86_64.sh
- Accept the license and default location, answer “yes” to initialize, then close and reopen the terminal.
- conda-forge is already the default channel, so no further channel setup is needed.
Step 5: Create the course environment (with Quarto and Jupyter)
Check first if you already have the
dsanenvironment: in Ubuntu runconda env list. Ifdsanis listed, skip creating it and just runconda activate dsan.
- Create everything in one environment:
conda create -n dsan python=3.12 jupyterlab quarto pandas numpy matplotlib
conda activate dsan
quarto check
- In VS Code (connected to WSL), select the
dsaninterpreter and thedsannotebook kernel. - Alternative Quarto install: the official Linux
.debat quarto.org/docs/download.
Step 6: R in Ubuntu (optional)
Check first if you already have R installed: in Ubuntu run
R --version. If you see a version, R is installed.
- For Python-first work you can skip R here and use native RStudio (Part 1) for R.
- To install R inside WSL:
sudo apt install -y r-base
- Use R through the VS Code R extension, or the
radianconsole. - The RStudio graphical app does not run cleanly inside WSL; for the RStudio experience, use native RStudio from Part 1.
Verify (WSL)
- In the Ubuntu terminal:
python --version
conda --version
quarto --version
- In VS Code (connected to WSL), open a
.ipynb, select thedsankernel, and run a cell.
Optional: Command Line and File Paths
Why this matters
- Tools like Conda and Quarto often run from a terminal, so a little navigation goes a long way.
- In VS Code, use the built-in terminal (Terminal menu, “New Terminal”) so your editor and command line share the same folder.
Core commands
pwd(macOS, Linux) orcdwith no argument (Windows): show the current folder.ls(macOS, Linux) ordir(Windows): list the folder’s contents.cd foldername: enter a folder;cd ..: go up one level.mkdir name: make a new folder.
Paths
- Absolute path: the full location from the root, for example
/Users/you/dsanorC:\Users\you\dsan. - Relative path: measured from where you are now, for example
data/file.csv. - Home shortcut:
~on macOS and Linux. - PATH (the environment variable): the list of folders the system searches for programs; this is why “add to PATH” during install matters.
You Are Ready
Before lecture, confirm all of the following:
- R and RStudio open and run.
- Conda works, and
conda activate dsansucceeds. - VS Code runs a notebook cell using the
dsankernel. quarto checkpasses.
Bring your laptop with these done. If a step failed, copy the exact error message and ask before class.