Ubuntu Data Science Setup Script

Overview

This repository provides a modular, profile-based bootstrap system for data science and development environments on:

• Ubuntu 24.04 LTS (native)
• WSL2 (Windows Subsystem for Linux)
• Ubuntu running inside Hyper-V

The project is designed for:

• Reproducible environments
• Safe, idempotent installation
• Optional profiles instead of installing everything by default

---

Supported Environments

Environment	Supported	Notes
Ubuntu 24.04 LTS	Yes	Primary target
WSL2 (Ubuntu)	Yes	Separate WSL-specific logic
Hyper-V Ubuntu VM	Yes	Uses same modular installer
Other Ubuntu versions	Limited	Not fully tested

---

Key Features

• Modular profile-based installation
• Idempotent scripts safe to re-run
• Separate handling for native Ubuntu and WSL
• Official Docker installation (Docker Engine + Compose plugin)
• VS Code-based tooling (including SQL support)
• Optional GPU support (native vs WSL separated)
• Lightweight defaults with opt-in components

---

Installation

Clone the Repository

git clone https://github.com/tomblanchard312/UbuntuDataScienceSetupScript.git
cd UbuntuDataScienceSetupScript

---

Usage

Native Ubuntu

./ubuntu_post_install.sh --profile base,python,viz

---

WSL

./wsl_post_install.sh --profile base,python --light

WSL requires systemd for some features. If not enabled, instructions will be provided by the script.

---

Hyper-V

1. Use the PowerShell script to provision the VM.
2. Run the Linux installer inside the VM.

.\ubuntu-hyper-v-setup.ps1 -VmName "DataScienceVM"

---

Profiles

Profiles allow you to install only what you need.

Profile	Description
base	Core system tools and dependencies
python	Python, pip, virtual environments, optional Miniforge
ml	Machine learning libraries
viz	Visualization and notebook tools
db	Database clients and VS Code SQL tooling
docker	Docker Engine and Compose plugin (official install)
k8s	Kubernetes CLI tools
r	R environment (optional)
gpu	GPU support (native Ubuntu only)
wsl	WSL-specific optimizations

---

Example Configurations

Minimal setup:

./ubuntu_post_install.sh --profile base,python

Data science workstation:

./ubuntu_post_install.sh --profile base,python,ml,viz

Full dev environment:

./ubuntu_post_install.sh --profile base,python,ml,viz,db,docker

WSL lightweight:

./wsl_post_install.sh --profile base,python --light

---

GPU Support

• Native Ubuntu GPU setup is handled separately.
• WSL GPU support relies on the Windows host NVIDIA driver.
• The scripts do not install Linux GPU drivers inside WSL.

---

Package Management Notes

• Uses modern APT repository handling (no apt-key).
• Uses official vendor repositories where appropriate.
• Docker is installed via Docker's official repository.
• VS Code is used for development and SQL tooling.

---

Python Environment Options

• Miniforge or micromamba is the recommended default.
• The installers default to Miniforge unless --with-anaconda is provided.
• Anaconda is optional.
• Select behavior with:
  • --with-miniforge
  • --with-anaconda

---

Idempotency

• Scripts can be safely re-run.
• Existing installations are detected and skipped where possible.
• The design supports repeatable environment setup.

---

Repository Structure

lib/        -> shared helpers
profiles/   -> modular install profiles
manifests/  -> package lists
docs/       -> supporting documentation

---

Documentation

• docs/architecture.md
• docs/wsl-notes.md
• docs/gpu-notes.md

---

Contributing

• Pull requests are welcome.
• Follow shell best practices.
• Ensure scripts remain idempotent.
• Test on Ubuntu 24.04 or WSL.

---

License

See LICENSE.