DATA 3402 — Python for Data Science 2

Fall 2026 · University of Texas at Arlington · Instructor: Amir Farbin

Course materials are distributed through this repository. Pull regularly — lectures and labs are pushed as the semester progresses.

Contents

Path	What's there
syllabus.pdf	Course syllabus, grading breakdown, policies
Lectures/	One folder per lecture: notebooks and/or slide PDFs
Labs/	One folder per lab assignment, plus the GitHub setup guide
sample.ipynb	Minimal notebook to check your environment works
requirements.txt	Python packages used in the course

Getting started

We do all of this together in the first lab session. These are the same steps, so you can catch up or start over.

You need Linux, macOS, or Windows with WSL. On Windows, WSL is required — not a suggestion. Google Colab is an acceptable fallback if your machine can't handle the later assignments, but it is not a substitute for having a working setup.

1. Get a Unix shell.

   • Windows: in PowerShell as Administrator, run wsl --install, then reboot. Launch Ubuntu from the Start menu and set a username and password when it asks. Write the password down — every sudo needs it.
   • macOS: open Terminal and run xcode-select --install.
   • Linux: you already have one.

2. Install the system packages. On Ubuntu/WSL/Debian:

   sudo apt update
   sudo apt install python3 python3-pip python3-venv git

   On macOS these came with the command line tools. Then close and reopen the terminal.

3. Make a course directory with its own virtual environment.

   cd ~
   mkdir -p Data-3402
   cd Data-3402
   python3 -m venv .venv
   source .venv/bin/activate

   Your prompt should now start with (.venv). That is a private Python for this course — nothing you install into it can break your system. Run the source line again in every new terminal; most "module not found" problems are just a forgotten activation.

   Do not use sudo pip install. Current Ubuntu blocks it with externally-managed-environment, on purpose; the virtual environment is the fix.

   Windows: keep all of this in your Linux home directory (~), not under /mnt/c. Git and Jupyter are slow and unreliable across that boundary. Your Windows drives are visible from Linux at /mnt/c, and your Linux files are visible from Windows File Explorer at \\wsl.localhost\Ubuntu.

4. Clone this repository and install the requirements. The clone is read-only for now; git pull brings you new lectures and labs as they are released.

   cd ~/Data-3402
   git clone https://github.com/UTA-DataScience/DATA3402.Fall.2026.git
   cd DATA3402.Fall.2026
   pip install -r requirements.txt

   No sudo on that last line — inside the virtual environment you are already allowed.

5. Check that it works.

   python3 -c "import numpy, pandas, matplotlib; print('ok')"
   jupyter lab

   Open sample.ipynb, click the code cell, and press Shift+Enter. Ctrl-C twice in the terminal stops Jupyter.

   Every time you come back: cd ~/Data-3402 && source .venv/bin/activate, then cd DATA3402.Fall.2026.

Notes on large data

Several lectures and labs use datasets too large to keep in git (the SUSY dataset, Kaggle competition data, image sets). Those are downloaded by the notebooks themselves and are excluded via .gitignore — don't commit them.

Labs and submission

Lab work is submitted through your own fork of this repository:

• you fork this repository once on GitHub, giving you your own copy under your account;
• you pull from this repository to receive new lectures and labs as they are released;
• you do your lab work in your fork and push to your fork, which is where it is graded.

Don't fork yet. The clone from Getting started is all you need to follow along, and it is deliberately read-only — you can pull, but you can't push to it. We set up forks together, step by step, in the lab session that covers git and GitHub, and that session reuses the directory you already have: the existing remote is renamed to upstream and your fork becomes origin. Nothing you do today has to be undone.

Labs/Git Hub Setup Guide (wsl & Mac Os).pdf is the written version of that session. Read it then, not now — and note that it still shows a previous semester's repository name in its example URLs.

Communication

All course communication goes through Teams — not email.