Skip to main content

Specialise in Data science

Data Science includes fields like machine learning, data mining, deep learning, artificial intelligence, optimization, and statistics, and it relates to terms such as data analytics and big data.

In many areas, incorporating insights from data analysis makes a major difference. Examples of data science in action in our everyday life are product recommendations in online stores or personal assistant systems on smartphones. Many companies want to use data science techniques to optimise their businesses. In the industry, machine learning, optimisation and artificial intelligence are applied in hot developing technologies such as robotics, drones, and self-driving cars.

Associate Professor Arthur Zimek in front of SDU's main entrance.

"Students just come by and ask, there is no barrier, no formality. It’s a friendly, supportive atmosphere. We are here to succeed together, to be a team in learning," says Arthur Zimek, Associate Professor in Computer Science.

In the Data Science group at SDU, statisticians and computer scientists work together for teaching, so we provide expertise in various aspects of our educational programmes as well as a coherent picture. Through close cooperation with other faculties, we are also able to offer courses that connect to the upcoming field of Personalised Medicine, a field which is relevant to everyone and which relies heavily on Data Science.

We are engaged in data science projects with various companies, from small and medium local companies to big players, which is why we can offer hands-on experience in student projects as well as theoretical research at the forefront of this field.

For instance, we are working with the City of Odense and the municipalities of Kolding, Nyborg and Svendborg on improving traffic systems, the planning of public transportation and the allocation of posts for building maintenance in the yearly budgets. In the industrial sector, we have current and past projects on data analysis and optimisation with Danfoss, Ørsted, Energinet, Lego and Aviation Cloud.

Courses offered

The Master's degree programme in Computer science at SDU allows you to choose most of your courses freely according to your personal interests. The following courses are the ones offered within the area of data science in the academic year 2019/2020:

DM847: Introduction to Bioinformatics

We will start with a concrete biological and/or medical question, transform it into a computational problem formulation, design a mathematical model, solve it, and finally derive and evaluate real-world answers from within the model. Students will be introduced to different computer science models and methods and their application within the area of Personalized Medicine, such as molecular biology, central aspects of gene regulation, epigenetic DNA modifications, and specialties with regards to bacteria & phage genetics

Responsible teacher: Richard Röttger

Read full course description

ST808: Multivariate Data Analysis and Chemometrics

This course teaches the competence to plan and execute scientific projects at a high level, including the management of work and development of situations that are complex, unpredictable and that require new problem-solving skills. Furthermore, it provides knowledge about advanced models and methods in applied mathematics, based on international research as well as the knowledge how to apply these models and methods on problems from various disciplines and the private sector.

Responsible teacher: Hans Christian Petersen

Read full course description

DM863: Deep learning

Machine learning has become a part in our everyday lives, from simple product recommendations to personal electronic assistants to self-driving cars. Especially Deep Learning has gained a lot of interest in the media and has demonstrated impressive results.

This intensive course will introduce students to the exciting world of deep learning. We will learn about the theoretical background and concepts driving deep learning and highlight and discuss the most noteworthy applications of deep learning but also their limitations. Furthermore, all content will be put into practice immediately by suitable exercises and programming tasks.

Responsible teacher: Richard Röttger

Read full course description

DM870: Data Mining and Machine Learning

Data Mining and Machine Learning techniques enable computational systems to identify meaningful patterns in the data and to adaptively improve their performance with experience accumulated from the observed data.

This course introduces the most common techniques for performing basic data mining and machine learning tasks, and covers the basic theory, algorithms, and applications. This course balances theory and practice, and covers the mathematical as well as the heuristic aspects. Computational learning methods are introduced at a general level, with their basic ideas and intuition.

Moreover, the students have the opportunity to experiment and apply data mining and machine learning techniques to selected problems.

Responsible teacher: Arthur Zimek

Read full course description

DM871: Linear and Integer Programming

The main focus of linear and integer programming is on resource constrained optimisation problems that can be described by means of linear inequalities and a linear objective function. These problems may arise in all contexts of decision making, such as manufacturing, logistics, health care, education, finance, energy supply and many others.

In this course, you will learn the basics of linear and integer programming and duality theory and the main solution techniques, such as the simplex method, branch and bound and cutting planes. The course also aims to provide hands-on experience with mathematical modeling and the solution of these models using software systems.

Responsible teacher: Marco Chiarandini

Read full course description

DM872: Mathematical optimization at work

The course is a continuation of DM559: Linear and Integer Programming and focuses on advanced solution techniques to tackle concrete applications from practice.

The course aims at giving the theory behind the solution techniques and above all at gaining practical experience in deploying them on a few concrete numerical instances for optimization problems taken from scheduling and vehicle routing applications.

Examples of such techniques are: delayed column generation, Lagrangian relaxation and Bender decomposition.

Responsible teacher: Marco Chiarandini

Read full course description

DM865: Heuristics and approximation algorithms

Many optimization problems from industrial applications such as scheduling, logistics, energy planning, sports planning, etc., can be formulated as discrete optimization problems, but they can not be resolved optimally within a reasonable time. Here, heuristics, metaheuristics and approximation algorithms play an important role.

General heuristics and metaheuristics are loosely defined rules to proceed to near-optimal solutions. They are often inspired by nature. For example, local search techniques are based on the principle of trial and error, which is a possible way in which humans intuitively solve problems.

Differently from heuristics, approximation algorithms come with a guaranteed running time and solution quality.

Responsible teacher: Marco Chiarandini

Read full course description

ST811: Multivariate statistical analysis

This course focuses on multivariate statistical methods used for dimensionality reduction, analysis of mean vectors, and discrimination and classification. Such methods are of relevance for a wide range of practical applications: quality control of industry machinery, epidemiology and clinical problems in population health care, and questions in biological conservation and environmental monitoring.

Responsible teacher: Hans Christian Petersen

Read full course description

ST813: Statistical Modelling

The aim of the course is to enable you to gain insight into the mathematical structure of linear and generalised linear models, including experience in recognising such models from a given statistical problem.

Responsible teacher: Fernando Colchero

Read full course description

ST816: Computational Statistics

The aim of the course is to enable the you to use modern computer-intensive statistical methods as tools to investigate stochastic phenomena and statistical procedures, and to perform statistical inference, which is important in regard to conducting statistical analysis based on computation and simulation.

Responsible teacher: Yuri Goegebeur

Read full course description

 

Most of these courses require programming abilities and a basic understanding of linear algebra. You can choose any number of the courses, as long as you meet any prerequisites, and as long as the courses fit into your course of study.

Master Thesis projects

The following are examples of previous Master Thesis topics in the area of Data Science:

  • Simulation of Traffic Flow in a Real Urban Network (2018)
  • Optimization of coordinated traffic signal intersections (2018)
  • Flight Planning in Free Route Airspaces (2017)
  • Artificial intelligence in action real-time strategy games (2016)
  • Optimizing heat and power production using column generation (2016)

Who teaches Data Science?

Marco Chiarandini is fascinated by the fantastic journey that optimizers undertake to solve timetabling, scheduling and routing problems. The journey moves forward in the abstract world through problem communication, mathematical representation, algorithm design, implementation and experimental analysis. Finally, it returns to the real world with numbers that correspond to actually practicable decisions, that yield systemic improvements and can ultimately ameliorate our lives.

Fernando Colchero is a statistician with particular interest in developing and applying inference methods to understanding population dynamics and demographic trends across the tree of life. He uses Bayesian inference as the statistical framework to explore hidden processes that affect and drive natural populations.

Lene Favrholdt likes to get to the core of an algorithmic problem. She strives to understand the essence of what results can be obtained, looking for precise and intuitive explanations of how and why they can (or cannot) be obtained. She finds communicating this understanding to colleagues and students in a clear and intuitive way very satisfying.

Yuri Goegebeur focuses on theoretical statistics, with main interests in extreme value theory, order statistics and asymptotic results. Although his work is largely theoretical, he has also applied extreme value methods to finance and insurance (claim size modelling, reinsurance), environmental science (ozone pollution and temperature), health science (modelling longevity) and geostatistics (earthquake magnitudes, diamond valuation).

Hans Christian Petersen is a biological anthropologist working in applied statistics, especially focusing on methods and applications related to human, primate and general evolution. Research topics are morphology, multivariate statistics and ways of dealing with datasets with missing values.

Jing Qin is a mathematician whose research areas cover combinatorics, graph algorithms and their applications in RNA Computational Biology. Most recently, she has started working in extreme statistics and applications. She enjoys applying mathematics and statistics to solvin real-life problems, although it sometimes means struggling in between theoretically neat results and not-so-neat practical applications. Well, all struggles lead to better scientific discoveries - eventually.

Richard Röttger is a computer scientist and specialises in various fields of Bioinformatics. In his research he focuses on the analysis of biological networks and large-scale biomedical datasets with the aim of utilizing existing information as efficiently as possible and extracting knowledge from the plethora of available biological data. To that end, he and his group employ state-of-the-art machine learning techniques like deep learning in order to research genes and proteins not in isolation but as a complex choreography of interactions leading to a deeper understanding of organisms and diseases.

Peter Schneider-Kamp

Arthur Zimek is not only a computer scientist by training but also studied philosophy, where one of the fundamental questions is: “what can we know?” This question translates into data science by the quest to understand intuition, based on which some machine learning methods work, what assumptions machine learning methods require, and how we can trust and interpret their results.