Analytical Methods

Visual and interactive analytics: transparency instead of  black boxes

We always focus on methods of analysis which give industry experts transparent results based on an intelligent combination of human know-how and machine analytics.  Visual analytics allows us to explore unknown data and to examine staticial models for plausibility. Our specialised tools support data scientists in their daily work and allow industry experts to apply their knowledge in assessing and interpreting the patterns they find.


Pattern recognition: tapping texts

Significant information is often hidden in unstructured data and texts making them difficult to analyse with traditional methods. Modern text mining technology, however, means even complex and unclear documents can be tapped automatically to extract specific hard facts, general topics or subjective emotions. In this way we can gain completely new insights, particularly from mass data such as  social media.and archives with  internal documentation.


Deep learning: tackling complex relationships

Significant patterns will very often not show up at the level of individual signals, transactions or customers but only when many different types of data are connected. It is only through smart processes that a deeper picture emerges: Graph mining can illustrate complex network dependencies, deep learning takes into account complex numerical patterns in large quantities of data and data stream mining operates on high-frequency data streams. We are intensively investigating innovative approaches so that you can rise to the challenge and benefit from big data.


Privacy-preserving data mining: data protection built-in

Personal data is strongly and specifically protected by law and must be treated correctly within the data analysis process. Third parties must not be able to recognise critical business data from analytic results. In both cases privacy-preserving data mining uses mathematics to guarantee the required degree of data protection.