Introduction

Data Science: What and why?

Data science is an interdisciplinary field about processes and systems to extract knowledge or insights from data. The goals of Data Science include discovering new phenomena or trends from data, enabling decisions based on facts derived from data, and communicating findings from data. It is a continuation of some of the data analysis fields such as statistics, data mining, and predictive analytics.

Data Science is at the heart of the scientific method, which starts with making data-driven observations to formulate testable hypotheses. It furthermore comes into play to visualize and assess experimental results. Data science skills are therefore necessary to any field of scientific research. Data science is the main tools of epidemiology, the study of health and disease in populations, which largely relies on observational data. Moreover data science is important in the industry, to understand operational process, and in business analytics, to understand a particular market. Hence, with the rise of big data in all areas of society, data science skills are some of the most demanded skills on the job market. Last, but not least, in an era of fake news, data science skills are important for citizens of modern societies.

Scope of the course

The goal of this course is to provide general analytic techniques to extract knowledge, patterns, and connections between samples and variables, and how to communicate these in an intuitive and clear way.

This course focuses on front-end data science. This means, it teaches practical skills to analyse data. We will focus on tidy data, visualizations, and data manipulation in Python. To only then dive into the math required to understand and interpret analysis results.

This course does not teach back-end data science, i.e., it does not teach how to develop a statistical or machine learning models, nor how to develop scalable data processing software.

Other courses offered by the faculty of Informatics cover data science back-end skills.

Python vs. R

Python is a versatile high-level programming language used that is widely used by the machine learning and scientifc programming community. Due to the increase use of machine leaning in data science, this course is taught in python.

An alternative language for Data Science is R. As of 2026, R typically offers more legible, compact and elegant code than python for front-end data science. Also many statistical methods have been natively developped in R with typically better default behaviour. Until 2025, this course used to be taught in R. The script is available at: https://gagneurlab.github.io/dataviz/

Course overview

The lecture is structured into three main parts covering the major steps of data analysis:

  1. Get the data: After basic introduction to Python, learn how to fetch and manipulate real-world datasets. How to structure them to most conveniently work with them (tidy data).

  2. Look at the data: Basic and advanced visualization techniques allows navigating large and complex datasets, identifying interesting signal, and formulating hypotheses. Typical sources of confounding are discussed. Recommendation to present an analysis in compelling fashion are also given.

  3. Conclude: Concepts of hypothesis testing will allow concluding about the statistical robustness of discovered associations. Also, methods from supervised learning will allow to model data and build accurate predictors.

The chapters of this script corresponds to individual lectures. Appendices provide further technical details as well as Python tricks and tips.

Complementary reading

These books offer complementary information to this script: