starter-data-python

A Skilluv starter — JupyterLab + pandas + DuckDB + scikit-learn, with a sample African dataset.

English

What this is

A ready-to-run Jupyter Lab environment for data exploration:

• JupyterLab 4 in Docker
• pandas 2 for data manipulation
• DuckDB 1 for SQL over CSVs and Parquet, no DB server needed
• scikit-learn 1.5 for classical ML
• matplotlib + seaborn for plots
• Managed with uv
• Sample dataset shipped: 2023 African population + GDP per capita (illustrative)

Quickstart

git clone git@github.com:skilluv-community/starter-data-python.git
cd starter-data-python
cp .env.example .env
docker compose up --build

Open http://localhost:8888 and browse notebooks/.

Notebooks

• 01_load_data.ipynb — load the CSV, inspect shape
• 02_basic_analysis.ipynb — descriptive stats + seaborn plots
• 03_duckdb_queries.ipynb — SQL over the CSV via DuckDB
• 04_ml_intro.ipynb — minimal scikit-learn regression

Docs

• docs/en/getting-started.md
• docs/en/architecture.md

---

Français

Environnement JupyterLab prêt-à-l'emploi pour l'exploration de données. Voir docs/fr/getting-started.md.

git clone git@github.com:skilluv-community/starter-data-python.git
cd starter-data-python
cp .env.example .env
docker compose up --build

---

License

MIT — see LICENSE.