The problem Git solves
Without Git, projects turn into folders like final.csv, final_v2.ipynb, and final_REALLY_FINAL.ipynb. Git gives you a timeline of changes, so every experiment has context.
What Git tracks
- Code - notebooks, Python scripts, SQL files, and configs.
- Decisions - commit messages explain what changed and why.
- Project state - you can return to an older working version.
Git vs GitHub
Git is the version-control tool on your machine. GitHub is a hosting and collaboration platform for Git repositories.
Essential commands
Why identity matters
Every commit stores an author name and email. This is how GitHub connects your local commits to your profile.
Core workflow
Good commit messages
A useful commit message says what changed in a specific, readable way. Use verbs like Add, Fix, Refactor, Document, and Update.
Common branch workflow
Naming branches
Use clear names such as feature/sales-dashboard, fix/readme-setup, or analysis/customer-churn-eda.
Local to GitHub
Daily sync commands
Files you usually ignore
.envfiles with API keys.- Large raw datasets and exports.
- Python cache folders like
__pycache__/. - Model binaries like
*.pklunless deliberately versioned elsewhere.
Recommended layout
This layout separates exploration, reusable code, data, and final reports. It also makes the repository easier to review.
Two common files
Better manual requirements for portfolio projects
A manually curated requirements file is often cleaner than freezing every package from a large environment.
README sections that matter
- Project goal and business question.
- Dataset source and limitations.
- Method summary.
- Key findings with visuals.
- How to run the project.
Release criteria
- Notebook runs top-to-bottom.
- README explains the business question and findings.
- No secrets or private data are committed.
- Charts are readable and decision-focused.
- Environment and data instructions are clear.