dMAT Data Science Module 2026: Syllabus, Topics, Practice Questions

Data Science is a Subject Module of the dMAT exam that applies only to Master's applicants targeting the M.Sc in Applied Data Science at Georg-August-Universität Göttingen. It has a duration of 90 minutes and measures your ability to apply data science fundamentals from a Bachelor’s degree to typical academic problems. 

This guide covers everything you need to know about the Data Science Module in the dMAT exam, including syllabus, preparation tips, and a study plan.

Key Highlights:

• The dMAT Data Science module lasts 90 minutes and does not permit calculators, notes, or rough sheets during the exam. 
• The dMAT Data Science question format uses single-choice questions based on a subject-specific passage, table, formula, or image. 
• The dMAT Data Science syllabus covers three key test areas: Theoretical Computer Science, Applied Computer Science, and Mathematics.

dMAT Data Science Module Overview

Data Science is one of seven dMAT Subject Modules and applies only to the Data Science degree programme at Georg-August-Universität Göttingen. While the dMAT Data Science Module is optional for Göttingen admission, a strong score strengthens your application in a competitive selection round.

Here is a quick overview of the dMAT Data Science module:

dMAT Data Science Syllabus

There is no fixed dMAT Data Science syllabus published. The areas of Theoretical Computer Science, Applied Computer Science, and Mathematics represent only a selection of the possible topics that may be covered in the test. The topics outlined below should therefore be used as a guide to the key areas to build fluency in, rather than as a complete checklist. 

Theoretical Computer Science

Theoretical Computer Science questions centre on combinational logic. You need to read and construct truth tables for AND, OR, and NOT gates, derive a Boolean function from a truth table using the canonical disjunctive normal form (CDNF), and work backwards from a circuit diagram to the truth table it produces. Questions typically frame these as a short real-world scenario, such as an alarm or sensor system, rather than as pure symbol manipulation.

Applied Computer Science

Applied Computer Science question in the dMAT Data Science Module tests how programming languages handle data at the variable level. Expect questions on choosing the correct data type (boolean, short, int, float, double, string) for a given value, how implicit type conversion works when two different types are used in one calculation, and how memory requirements scale with data type and array size. Familiarity with Java’s type rules, rather than just general programming logic, is particularly useful.

Mathematics

The Mathematics question draws on linear algebra concepts applied in a data science context. You should be able to calculate the rank of a small matrix, find eigenvalues and eigenvectors, compute a dot product, and reason through what Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) do to the dimensionality of a data set. Several questions build on the same base matrix across a sequence, so an early calculation error compounds through later questions in that set.

Expert Insight: While the dMAT Data Science syllabus overlaps with the dMAT Computer Science module, the two modules do not follow the same task format. The Data Science module does not have a published Basic Task and Advanced Task structure. You should therefore focus on applying core concepts across the stated areas rather than preparing for separate task categories based on difficulty. 

dMAT Data Science Practice Questions

The three Data Science practice questions below reflect the structure of official dMAT exam questions. Attempt them without notes or a calculator, as you would in the actual exam.

Sample Question 1: Applied Computer Science

In Java, several data types are commonly used to represent different kinds of values. The following are examples:

• boolean: Represents true or false values and requires 1 bit of memory.
• short: Stores integer values from −2¹⁵ to 2¹⁵ −1 and requires 16 bits of memory.
• int: Stores integer values from −2³¹ to 2³¹ −1 and requires 32 bits of memory.
• float: Stores floating-point numbers with values of up to approximately ±10³⁸ and requires 32 bits of memory.
• double: Stores floating-point numbers with values of up to approximately ±10³⁰⁸ and requires 64 bits of memory.
• string: Stores a sequence of characters of any length, with memory usage depending on the length of the string.

A data pipeline needs three variables: recordCount, which stores how many rows a batch script has processed so far and never exceeds 30,000; isValidated, which stores whether the batch has passed a data quality check, and pipelineName, which stores the name of the pipeline. Which combination of Java data types fits these three variables most efficiently?

1. a) short recordCount; boolean isValidated; string pipelineName
2. b) int recordCount; float isValidated; string pipelineName
3. c) boolean recordCount; short isValidated; string pipelineName
4. d) float recordCount; int isValidated; string pipelineName

Answer: a. A short in Java covers integers from roughly -32,768 to 32,767, which comfortably fits a count below 30,000 while using only 16 bits instead of the 32 bits an int would need. isValidated is a yes-or-no value, so boolean (1 bit) is correct. pipelineName is a sequence of characters, so string is correct. Options b, c, and d each assign a type built for one kind of value to a different kind, such as a boolean for a numeric count.

Sample Question 2: Theoretical Computer Science

Remember the Boolean operations AND, OR, and NOT. Their representation as logical gates in circuits is shown in the figure below.

They are represented formally as functions 𝑋 = (𝐴 ∧ 𝐵) for AND, 𝑋 = (𝐴 ∧ 𝐵) for OR and 𝑋 = 𝐴 for NOT. The output X can be given by the combinations of the possible inputs. For Boolean operations, the input can only be 0 or 1 (sometimes also referred to as false and true). For completeness, the output of all three operations is presented using truth tables: 

From a truth table, we can immediately derive a Boolean function or the combinatorial circuit by using what is called the canonical disjunctive normal form (CDNF). For each 1 that results from a function, a minterm is created. A minterm is an AND operation over all inputs, where 0-inputs are negated while 1-inputs are not. 

All minterms are finally combined with an OR operation. As an example, the CDNF of AND derived from its truth table has only one minterm 𝑋 = (𝐴 ∧ 𝐵). It thus does not need to be used in an OR. However, the OR operation as CDNF has three minterms that need to be combined with OR: 𝑋 = (𝐴 ∧ 𝐵) ∨ (𝐴 ∧ 𝐵) ∨ (𝐴 ∧ 𝐵).

A data quality alert fires if either of two checks fails on an incoming file: a missing value check (input A, where 1 means the check failed) or a duplicate record check (input B, where 1 means the check failed). Which Boolean operation correctly describes the alert output, X?

a) AND
b) OR
c) NOT
d) One AND combined with one OR

Answer: b. The alert should fire the moment at least one of the two checks fails. That behaviour matches the OR truth table exactly: the output is 1 whenever A is 1, B is 1, or both, and 0 only when both A and B are 0. An AND gate would only fire the alert when both checks failed at once, which misses a file that fails just one check.

Sample Question 3: Mathematics

An eigenvalue is a scalar value that represents how the linear transformation stretches or shrinks a vector in a given direction. To calculate the eigenvalues of a matrix, solve the equation: det(A − λI) = 0, where A is the matrix, λ is the eigenvalue, and I is the identity matrix of the same size as A. The determinant det of a matrix is a scalar value that indicates whether the matrix is invertible and describes the scaling factor of the linear transformation represented by the matrix. For a 2x2 matrix, 𝐴 = ( 𝑎 𝑏 𝑐 𝑑 ), the determinant is calculated as: det(A) = ad – bc. 

Consider the matrix below

| 4 0 |

| 0 2 |

What are the eigenvalues of A?

a) 4 and 2
b) 6 and 8
c) 2 and 2
d) 4 and 4

Answer: a. For a diagonal matrix, where every entry off the main diagonal is zero, the eigenvalues are always equal to the diagonal entries themselves. You can verify this by solving det(A − λI) = 0: (4 − λ)(2 − λ) − 0 = 0 gives λ = 4 and λ = 2.

4-Week Study Plan for the dMAT Data Science Module

A four-week dMAT Data Science preparation should be split across the three test areas. The focus should be on building conceptual understanding, applying data science principles to problems, and improving speed and accuracy through timed practice.

How to Prepare for the dMAT Data Science Module?

Preparation for the dMAT Data Science module should focus on understanding the tested concepts, applying them accurately, and building speed through timed practice. 

• Build Java fluency even if you primarily code in Python or another language: Applied Computer Science questions rely on Java-specific type rules, so familiarity with Java’s type system is essential for avoiding errors in these questions.
• Practise without a calculator from day one: Since calculators are not allowed in the exam, practising eigenvalue and dot product calculations mentally from the start helps build speed for exam day.
• Do not skip Core Module practice: The Subject Module comes after a 75-minute dMAT Core Module in the same exam session, so treat your preparation as one combined three-hour exam rather than two separate tests.

From the Desk of Yocket

The dMAT Data Science module can be a valuable part of a Göttingen application, but effective preparation requires targeted practice across its three test areas rather than broad data science revision. 

Yocket Prep offers structured practice resources to help you build speed and accuracy across your Germany Master's preparation. With focused preparation and consistent practice, you can approach the test with greater confidence and readiness.