Data Science and Engineering (DSE)

Our Data Science and Engineering (DSE) research group designs and implements applications for the recording, integration, management, search and analysis of data.

Our research group

We develop data science applications for structured and unstructured data using cutting-edge methods (mathematics, statistics, machine learning), processes (Scrum, Design Thinking) and technologies (Spark, Tensorflow, R). Our expertise covers a wide range of areas of application (CRM, administration, GIS).

Range of services

In cooperation with national and international partners from industry, public administration and universities, we carry out application-led research projects on data-oriented applications:

  • Concept formulation and architecture design
  • Consulting and expert opinions on technologies and business cases
  • Modelling and consulting with regard to business processes
  • Design, development and evaluation of data analysis models
  • Implementation of application prototypes
  • Provision of individualised training

Areas of expertise

Modelling and analysis of structured data

The storage and management of structured data in an SQL database requires a carefully designed data model, as subsequent modifications usually involve significant cost and effort. With our many years of experience, we would be pleased to advise you on modelling. We can also support you with the development of complex projects and evaluations.

Analysis of unstructured data (text analysis, NLP)

We develop solutions for the automated analysis of texts (web, social media, e-mails, documents). This enables targeted information to be obtained (“Which products are mentioned?”), texts to be classified (“Is the e-mail an order or a support request?”) and compared (“Which other e-mails refer to the same problem?”).

Linked data, open data and semantic web

Linked data or ontologies allow knowledge to be formalised and evaluated, for example, as the basis for text analysis, semantic search or expert systems (decision-making support). This kind of (and other) data is also increasingly available publicly (open data).

Agile development of prototypes

We use agile methods – such as Design Thinking and Scrum – when developing new applications. This allows us to design key functions efficiently, to test out alternatives and to develop meaningful prototypes.

Geographical information systems (GIS)

The analysis and visualisation of geographical data generate new information and intuitively open up accessible insights into questions concerning geography. We support the design and implementation of geographical data analyses and the set-up of related databases and geographic information systems based on Esri ArcGIS or open source GIS solutions.

Machine Learning

Starting from business value and use cases, we rely on machine learning to develop practical solutions for data analysis and the management of smart applications. We have extensive experience in various fields of industry (legal, banking, healthcare, public administration), we work with different types of data sources (relational, document and graph databases) and leverage state-of-the-art methods (deep learning, transfer learning) and technologies (TensorFlow, Apache Spark). 

Augmented Intelligence

Automated learning and data analysis play an increasing role in offering innovative services and products. We offer our know-how in various fields of Artificial Intelligence and examine how such technologies can be embedded in existing workflows. How can humans work most effectively with Artificial Intelligence? In what way can their respective skills complete one another?

Energy-efficient and scalable cloud and container platforms

To run smart applications, we need scalable and robust platforms. It is essential to reach a strong degree of automation for the application deployment processes and to run intelligent platforms between data centers and private and public cloud. In the field of Green IT, we develop models allowing the energy consumption in the data center to be measured and optimised.

Fairness and digital ethics

Digital ethics, in particular with regards to machine learning, become increasingly relevant. If the training data is biased, this can have an impact on the resulting models and lead to unfair decisions being made by the system. We develop methods to measure fairness, in particular in text data, and to apply digital methods to social and community problems.  

Contact

Contact us or meet our experts in person at various events. Collaboration produces win-win outcomes for everyone concerned – your company, society and the university.