Data Science (DS)

The way in which data is available and in which form has changed considerably with the widespread use of social media, extensive log data and so-called value-added services. What they all have in common is that customer data represents the main value. Examples include well-known social media representatives, payback card systems, health trackers, GSM and GPS trackers in cars and mobile phones, Internet of Things data and tracking systems for user click behaviour on the Web.

The task of data science is, on the one hand, the infrastructural challenge of dealing with the mostly extensive data, some of which arrive as a data stream. On the other hand, however, it is necessary to generate value from this data in the form of correlations. This is the actual core task. The resulting data is often unstructured and the correlations to be found are partly in the area of previously unnoticed questions, for which there are no preconceived solutions. Therefore, data mining methods used include unmonitored, but also monitored machine learning. This happens, for example, in the form of weak AIs on the basis of a deep learning architecture. Typical statistical topics here are regression, categorization, cluster formation/similarities and survival models.

---

Contact: Arne Scheffer (CIT)
Fon: 0251 83-31581