INF34306 Concepts in Data Science

Notebooks and datasets for Concepts in Data Science practicals.

Using the notebooks

The notebooks are meant to be used in Google Colab. Open a notebook with one of the links below, then use File → Save a copy in Drive to keep your changes.

Block	Tutorial	Project
1 - Introduction
2 - Python libraries, Visualization, and the Grammar of Graphics
3 - Data exploration I
4 - Data exploration II

In blocks 1 and 2, students download the datasets from Kaggle themselves and upload them to Colab (files.upload()). From block 3 onwards, the notebooks read datasets directly from this repository on GitHub (DATA_URL in the notebooks). Note that this points to the main branch: files that are not on main yet will not load.

Repository structure

Notebooks with exercises and examples are in the notebooks folder, with one subfolder per block. Datasets (when not available through APIs) are in the datasets folder, with informative filenames.

|-- README.md
|-- TODO.md
|-- datasets
|   |-- UniComplaints.csv
|   |-- anscombe_quartet.csv
|   |-- automobile.csv
|   |-- breast-cancer.csv
|   |-- iris.csv
|   |-- titanic.csv
|   |-- top_1000_popular_movies_tmdb.csv
|   `-- ... (not used by the current notebooks, see TODO.md)
|-- notebooks
|   |-- block1_introduction
|   |   |-- tutorial_1.ipynb
|   |   `-- project_1.ipynb
|   |-- block2_visualisation
|   |   |-- tutorial_2.ipynb
|   |   `-- project_2.ipynb
|   |-- block3_exploration_1
|   |   |-- tutorial_3.ipynb
|   |   `-- project_3.ipynb
|   `-- block4_exploration_2
|       |-- tutorial_4.ipynb
|       `-- project_4.ipynb
|-- pyproject.toml
|-- requirements.txt
`-- src
    `-- data_science_concepts   (course package, see below)

Images used in the notebooks are stored next to the notebooks and linked by their GitHub URL, because Colab does not resolve relative image paths.

Course package

Helper functions used in the notebooks (e.g. plot_decision_boundary) are in the data-science-concepts package in src/data_science_concepts. The notebooks install it with

%pip install -q "data-science-concepts==2026.*"
import data_science_concepts as dsc

To release a new version, create a GitHub release with a tag of the form vYEAR.MINOR.PATCH (e.g. v2026.0.1). The Package workflow (.github/workflows/package.yml) builds the package and publishes it to PyPI; the version number is taken from the tag. Notebooks pin the course year, so a release for the next year (v2027.0.0) does not affect notebooks of the current year.