Section for Statistics and Data Analysis

The section conducts research across a broad spectrum of pure and applied mathematics ranging from analysis of stochastic processes to development of computational methods in machine learning and artificial intelligence.

In the digital age of society, statistics and data analysis play an increasingly important role in transforming data into knowledge. The rapid growth in data availability creates new technological and economic opportunities, but also fundamental challenges.

 

The Statistics and Data Analysis section at DTU Compute develops novel methods and designs innovative solutions addressing these complex challenges. 


By advancing and combining theoretical and applied research in mathematics, we tackle pressing societal issues related to climate, public health, and equality.


Our research involves statistical analysis of various data modalities, for example images, text, and audio, with applications across science and engineering in diverse fields such as healthcare, finance, educational technology, and manufacturing.

 

Close collaborations with hospitals, multinational companies, and research institutions around the world ensure that our work remains scientifically rigorous and practically relevant.

 

The section conducts research across a broad spectrum of pure and applied mathematics ranging from analysis of stochastic processes to development of computational methods in machine learning and artificial intelligence. Our work can broadly speaking be categorized into four core research areas.

 

Education


We are responsible for teaching the mandatory introductory courses in statistics and offer advanced graduate-level courses in areas such as stochastic processes, computational data analysis, and statistical modelling and quality control. 

Across all levels, we emphasise scientific reasoning and critical thinking, providing a solid foundation for students in the Applied Mathematics (BSc) and Mathematical Modelling and Computation (MSc) programmes.

 

Data Insights Team

 

The Data Insights Team, DIT, constitutes a dedicated function within the Statistics Section, focusing on the application and dissemination of data analytical methods.

DIT is a research-based group working with data analysis, statistics, machine learning and AI across a broad range of application areas, including sectors such as construction, food, production and healthcare, as well as technologies like drones, translating advanced methodology into practical solutions.


The team collaborates widely with companies, public authorities, and colleagues across DTU, with the aim of strengthening data-driven methods and digital competencies across disciplines and sectors. 

 

 

Head of Section

Bo Friis Nielsen

Bo Friis Nielsen Phone: +45 45253397

Research areas

Applied Probability

We formulate and apply stochastic models for dynamical systems in a variety of fields. By combining theoretical analysis with numerical simulation, our research enables parameter estimation and statistical testing in complex multivariate models. Representative examples include modelling delays in railway systems, bed occupancy rates in hospital wards, and insurance risks.   

 

Lead researcher

Bo Friis Nielsen Professor, Head of Section

Modern Statistical Models

Modern computing facilities enable us to analyse vast amounts of data from our society using statistical learning and statistical engineering. The research focuses on regularisation strategies, bias mitigation, and sparse methods, and we apply deep neural networks and similar computational methods from machine learning within domains like psychiatry, chemistry, and genetics.

Lead researcher

Line Katrine Harder Clemmensen Associate Professor

Statistical Design and Analysis of Experiments

Experiments in manufacturing processes and clinical trials in the pharmaceutical industry are typically expensive and time-consuming. Rigorous planning based on statistical analysis and optimisation is therefore indispensable for efficiently obtaining reliable test results, thereby reducing production costs and increasing consumer safety.

Lead researcher

Murat Külahci Professor

Statistical Process Control

In collaboration with leading industrial partners, we design control schemes to monitor and improve industrial products and processes. Analysing time series and process data obtained from sensors, we decompose the variability in data to identify issues such as material impurities and defective machine components and to prescribe corrective actions.

Lead researcher

Murat Külahci Professor

Discover our research

Want to know more about our research? Visit our page in DTU's research database.
DTU Orbit

Staff

Subpages