Logo von VRVis Logo von VRVis

Research topics

Data Science

The common goal of research in data science is to extract knowledge, produce insights and solve problems with data. Data science systems are based on models and algorithms that are created to detect valuable patterns in the data and, through learning, to enable its predictive power. At VRVis we approach data empirically and research it with state-of-the-art competence for the purpose of practical applications in science, industry or technology.

Main research areas

Graphical representation of the data science ecosystem

From data preparation to analysis - at VRVis we are familiar with the Data Science pipeline* and are researching new methods and applications that optimally complement this pipeline.

  • Discover: Find the right data for your task. This is often not easy, as data is stored in different databases and repositories in many organizations. In addition, the selected data must fit the initial question.
  • Wrangle: Parse the data and bring it into the right format. The "Wrangle" step is at the same time the most unpleasant, but also the most time consuming step. In this step the data must be put into a format so that further work can be done with it. Then includes parsing the data, dealing with different data formats, handling missing and erroneous data, and merging different data sources.
  • Profile: Explore the data, find interesting patterns and features, with the use of interactive visualization tools. This step is about getting to know the data better, analyzing its structure, and finding hidden patterns - the paradigm of visual analytics, the interactive, visual representation of data. With our visual analytics tools we can help out in an optimal way to provide users with a comprehensive insight into complex data.

  • Model: Build an abstract representation (a model) of the data - this can be an AI model, but can also be a simpler representation like a categorization, clustering, regression. Models of the data tell us more about the structure of the data and allow us to make predictions. In our Artificial Intelligence research we focus on teaching computer systems by machine learning or deep learning based on large volumes of training data to behave intelligently, analyze their environment and take decisions autonomously.

  • Report: Show your findings to other stakeholders, often by the use of interactive dashboards. New insights from data usually need to be communicated further, usually to people who are not involved in the analysis from the beginning, and therefore do not have the same insight into the data. Visualizations are of great advantage here, as they provide quick access to facts and insights.

* See Data Science Flow by: Kandel, S., Paepcke, A., Hellerstein, J.M., and Heer, J. (2012). Enterprise Data Analysis and Visualization: An Interview Study. IEEE Transactions on Visualization and Computer Graphics 18(12): 2917-2926

CONTACT

Dipl.-Math.in Dr.in Katja Bühler

science(at)vrvis.at+43 1 908 98 92

Reference projects, applications and publications

Screenshot of the Brain* software of the VRVis showing a gray brain with many colored neuronal networks.
The VRVis data science platforms not only make it easy to handle extremely large heterogeneous data, as is the case in life sciences and medicine, but also enable better data and data network comprehensibility through integrated visualization.

Data Science for Medicine and Life Science

Data Science plays an increasing role in the fields of medicine and life sciences. VRVis has already successfully implemented several data science projects in the areas of neuroscience and medicine for the analysis of health and patient cohort data. Furthermore, VRVis has developed Brain*, a globally used platform for the integration, management and exploration of spatial brain data.

Three people are standing in front of a laptop in a factory environment.
Data analysis tools optimized visually and graphically for human perception are the decisive key element that strengthens users' confidence in computer-aided data science and decision-making processes.

Visual Analytics for Industry 4.0

As a key technology in the field of industrial production, visual analytics enables us to see through the data surrounding the production process through visual data analysis. In this context, the stability and scalability of visual analytics tools are crucial factors for their acceptance as reliable tools in Industry 4.0, such as in our INGRESS project. Our analytics solutions are characterized by their ability to handle complex data, as well as to perform pattern search and anomaly detection quickly and reliably.

Bild einer Fischwanderhilfe und eines Staudamms mit Landschaft
The energy sector benefits significantly from intelligent visual data science solutions that contribute to the sustainable use and maintenance of hydropower plants, for example.

Digitalization and Predictive Maintenance in Hydropower

Hydropower is an important part of Austrian renewable energy. However, many hydropower plants are older structures. In our DIGI-Hydro project, a hydropower plant is retrofitted with new measurement sensors and advanced data analytics to facilitate real-time monitoring and better maintenance with the help of a digital (hydropower) twin.

Publications

Data Science

Awards

Projects