Die Forschungsgruppe "Business Data Analytics" setzt den Fokus auf die Chancen und Herausforderungen einer Welt, die immer stärker durch digitale Daten beeinflusst wird. Unser Ziel ist dabei stets handfeste Mehrwerte aus Daten und Informationen zu gewinnen und nutzbar zu machen. Dazu bedarf es einer ganzheitlichen technischen und ökonomischen Betrachtung der Anwendungsfälle. Wir forschen sowohl an Analytics/AI Methoden und Verfahren, ihrer praktischen Anwendung und der Einbindung in den Unternehmens- oder gesellschaftlichen Kontext. Zu unseren Stärken auf der technischen Seite zählen datengetriebene Analyse und KI, Maschinelles Lernen, und Hybridisierung von Methoden. Auf der wirtschaftlichen Seite haben wir Kompetenz in der Business Case Identifizierung, der Geschäftsmodellentwicklung, und im Marktdesign aufgebaut. Wir arbeiten branchenübergreifend in Industrie- und öffentlich-geförderten Projekten mit vielen Partnern erfolgreich zusammen.

Head of Area

Wolfgang Badewitz +49 (721) 9654 823  badewitz∂fzi.de

Members

	Tobias Kölbel   tobias.koelbel∂kit.edu		Patrick Jaquart +49 (721) 608-48387 patrick.jaquart∂kit.edu		Alexander Grote +49 (721) 608-48344  alexander.grote∂kit.edu
	Felix Sterk +49 (721) 608-48387  felix.sterk∂kit.edu

Domains

• Automotive
• Health Care
• Industry 4.0
• Mobility
• Public Management
• Telecommunications

Methods

• Statistical Learning
• Combinatorial Optimization
• Data Dimensionality Reduction
• Feature Extraction and Generation
• Incentive Engineering
• Data Fusion
• Method Hybridization

Projects:

• Responsible News Recommender Systems (ReNewRS)
   
• OCROSS
   
• Innovative Analytics in der Telekommunikation
   
• ReKoNet

Former Projects:

• Self-Sovereign & Trustworthy Marketplaces
   
• Data Quality Management im Corporate Financial Controlling
   
• Accuracy improvement of vast amounts of heterogeneous judgmental cash flow forecasts using analytical debiasing methods and combination with model forecasts (in collaboration with Bayer AG)
   
• Concise representation of cash flow forecasting- and revisioning-behavior processing analytical-orthogonal and Bayes-based metrics in corporate financial controlling (in collaboration with Bayer AG)
   
• Non-addictive Information Systems
   
• Design of robust and concise metrics to represent and cluster the purchasing and usage history of telecommunication customers used in campaign management (in collaboration with a global telecommunications company)
   
• Techniques to decompose constraint matrices in packing problems and step-wise generation of variance-preserving, pseudo-perpendicular constraints aimed at transforming high-dimensional MIP into lower-dimensional problem representations that allow for more efficient and scalable problem solving (in cooperation with Siemens AG)
   
• Productive 4.0
   
• Accuracy improvement of vast amounts of heterogeneous judgmental cash flow forecasts using analytical debiasing methods and combination with model forecasts (in collaboration with Bayer AG)
   
• Concise representation of cash flow forecasting- and revisioning-behavior processing analytical-orthogonal and Bayes-based metrics in corporate financial controlling (in collaboration with Bayer AG)
   
• Modelling and prediction of user behavior related electric vehicle high-voltage battery aging, based on heterogeneous field-data (in collaboration with a large German OEM)
   
• Design of robust and concise metrics to represent and cluster the purchasing and usage history of telecommunication customers used in campaign management (in collaboration with a global telecommunications company)
   
• Development of novel analytical approaches in the context of Geographic Information Systems (GIS) that allow a faster and more reliable consideration of vast amounts of heterogeneous and unreliable data in disaster and emergency management (BMBF-founded Project; program: Big Data
   
• Techniques to decompose constraint matrices in packing problems and step-wise generation of variance-preserving, pseudo-perpendicular constraints aimed at transforming high-dimensional MIP into lower-dimensional problem representations that allow for more efficient and scalable problem solving (in cooperation with Siemens AG)