Critical Thinking in Data Science with Python

From Observation to Responsible Decision-Making
by Mathias Ellmann

Book cover: Critical Thinking in Data Science with Python

ISBN: 978-3-695-26289-2

A Book About Data, Models, Judgment, and Responsibility

Data Science is often understood primarily as a technical discipline. Yet data do not speak for themselves. They must be collected, selected, structured, analyzed, interpreted, and connected to responsible decisions.

Critical Thinking in Data Science with Python presents Data Science as well-founded reasoning with data rather than computation alone.

The book brings together Data Science, statistics, epistemology, argumentation theory, decision science, fairness, transparency, communication, and human responsibility.

Python serves as a practical tool for exploring data, making reasoning processes transparent, testing hypotheses, and evaluating models without replacing human judgment.

Presentation Coming Soon Request a Workshop Buy the Book

Who This Book Is For

The book is intended for students, educators, Data Science beginners, Python learners, Data Analysts, Data Scientists, software developers, managers, decision-makers, project teams, IT consultants, and anyone who wants to understand and use data critically and responsibly.

Buy the Book

The English-language eBook edition of Critical Thinking in Data Science with Python is available from Amazon, Apple Books, Thalia, Hugendubel, and eBook.de.

Amazon Apple Books Thalia Hugendubel eBook.de

Topics and Key Themes

Reality, Data, and Observation

Why data are representations rather than reality itself, how measurement frames shape what becomes visible, and why observation, description, and explanation must remain distinct.

Facts, Interpretations, and Evaluations

How factual claims differ from interpretations and evaluations, and why conflating these levels leads to weak arguments and poor decisions.

Hypotheses, Correlation, and Causation

How patterns give rise to hypotheses, why correlation is not automatically causation, and how uncertainty limits the conclusions that data support.

Reasoning with Data

Deductive, inductive, and abductive arguments, along with common fallacies and cognitive biases that can distort Data Science.

Python as a Tool for Judgment

Using pandas, descriptive statistics, visualizations, and hypothesis tests to make analytical reasoning transparent and reproducible.

Critically Evaluating Models

Target variables, train-test splits, confusion matrices, ROC, AUC, overfitting, and the difference between strong metrics and sound decisions.

Decisions Under Uncertainty

Applying PrOACT, comparing alternatives, assessing risks, and justifying recommendations rather than treating model outputs as decisions.

Bias, Fairness, and Responsibility

Bias before and within models, competing fairness criteria, transparency, traceability, trust, automation, accountability, and human responsibility.

Communicating Data Science Clearly

The Information Iceberg—point, information, meaning, and intention—as a framework for precise, transparent, and responsible communication.

Presentation Coming Soon

An English-language presentation accompanying Critical Thinking in Data Science with Python is currently being prepared.

It will introduce data and reality, facts and interpretations, hypotheses, argumentation, uncertainty, Python, model evaluation, PrOACT, bias and fairness, human responsibility, and the Information Iceberg.

Frequently Asked Questions

What is Critical Thinking in Data Science with Python about?

The book presents Data Science as well-founded reasoning with data and connects observation, interpretation, hypotheses, argumentation, models, uncertainty, decisions, fairness, communication, and responsibility.

Is the book suitable for Data Science beginners?

Yes. It is intended for students, educators, beginners, Python learners, practitioners, and decision-makers who want to understand Data Science critically and responsibly.

Why is critical thinking important in Data Science?

It helps people examine assumptions, distinguish observations from explanations, recognize uncertainty, avoid reasoning errors, evaluate models appropriately, and justify decisions responsibly.

How is Python used in the book?

Python is used for understanding data with pandas, reading descriptive statistics critically, creating visualizations, testing initial hypotheses, and evaluating models.

What role does PrOACT play in the book?

PrOACT makes problems, objectives, alternatives, consequences, trade-offs, uncertainty, risk tolerance, and linked decisions explicit.

What is the Information Iceberg?

It is a communication framework that distinguishes point, information, meaning, and intention, helping analytical communication become clearer and more responsible.

Workshops and Lectures Based on the Book

The content can be adapted as a lecture, workshop, or moderated reflection format for universities, companies, educational institutions, Data Science teams, managers, and organizations.

The focus is on practical questions: What can data actually show? How can hypotheses and models be evaluated critically? How should uncertainty be communicated? How can alternatives and risks be compared? And where does human responsibility remain when analytical systems support decisions?

Request a Workshop Get in Touch

Contact

For enquiries about books, lectures, training programmes, or professional collaboration:

mail@mathiasellmann.de