Guided building and selection of decision tree models.

Efficient overview of data quality problems in many energy time series.

Detailed analysis of manufacturing data: Distribution of a quality indicator.

Detailed comparison of four surrogate models in automotive engineering.

The goal of the applied research project En2VA (“Visual Analytics for Energy and Engineering Applications”) is to increase the efficiency and the quality of advanced analytics for high-dimensional data from manufacturing, engineering, and the energy sector.

All these application domains are confronted with a rapidly growing number of data dimensions. Examples include smart-grid data in the energy context and an explosion of sensors in manufacturing. This often means a disruptive change to previous analysis methods and limits the applicability of existing standard tools for monitoring and analysis.

For such high-dimensional data, the project En2VA designs and develops new Visual Analytics methodologies for multiple important benefits:

  • Save time of domain experts and data scientists for data preparation by novel interactive approaches to data editing.
  • Build more accurate and understandable predictive models with higher confidence (e.g., regression models and decision trees) by novel visual approaches for model selection and validation.
  • Increase the confidence of decision makers based on new methods for automatically generating intelligent reports.
  • En2VA will also address scalability issues for creating effective overviews of high-dimensional data such as thousands of measured quantities as well as enable a progressive analysis of up to billions of data records.

In close collaboration with the company partners Austrian Power Grid, AVL List, HAKOM Solutions, Plasmo, and RHI, En2VA conducts a research approach with a strong focus on solving real-world problems of real users on real data. Specifically, all research results of En2VA are delivered as operational dashboards based on the software platform Visplore.