Project description

The aim of the Data Tales research project is to develop a hybrid AI approach to generative data storytelling and to apply it in the field of auditing. The central research question centres on the development of AI-supported methods for the automated generation of data stories by combining large language models (LLMs) with domain knowledge and interactive machine learning. In doing so, we are pursuing an innovative strategy that encompasses both collaborative interaction between humans and AI and the autonomous generation of data stories. The core of this approach lies in the ability to create coherent, informative and engaging narratives from extensive datasets and domain knowledge, which appeal to both subject matter experts and business users, thereby making complex data analyses widely accessible and understandable. This is intended to enable users to gain deeper insights into data through dialogue with a chatbot, and to refine and validate a narrative step by step. On the application side, we aim to sustainably improve the efficiency and quality of auditing by providing auditors with advanced methods that encourage a more in-depth examination of the audit results, thereby enabling them to focus on assessing the relevance and implications of the analysed data.

Research priorities at OTH Regensburg:

• Investigation of semantic methods for knowledge representation and their combination with large language models.
• Semantic representation of background knowledge in formal ontologies for data analysis processes and data storytelling.
• Development of machine learning methods for data storytelling based on LLMs in combination with knowledge-based methods such as ILP.
• Combining explainable interactive machine learning methods with intelligent dashboards and a human-in-the-loop approach.
• Research into innovative RAG approaches based on agents and language models.
• Conducting studies and experiments as part of a usability study to evaluate the research findings.

Student projects

We are always on the lookout for students who would like to contribute to the Data Tales project, whether through a dissertation, a project or as a student assistant. If you are interested, please do get in touch.

The project in brief

Title:

Data Tales – Generative and interactive data storytelling using generative artificial intelligence and knowledge-based systems

 

Collaboration partner:

• Emanuel Slany, Daten - Analysen & Beratung GmbH (DAB)
• Moritz Lang, Daten - Analysen & Beratung GmbH (DAB)
• Martin Riedl, Daten - Analysen & Beratung GmbH (DAB)
• Tassilo Föhr, (R&P)
• Frank Reutter, Rödl & Partner (R&P)
• Martin Wambach, Rödl & Partner (R&P)

 

 

Staff member (OTH Regensburg):

Sara Buchmann, MSc

 

Project Manager (OTH Regensburg):

Prof. Dr Stephan Scheele

 

Funding:

StMWi, BayVFP ‘Digitalisation’ funding programme

 

Duration:

November 2024 – October 2027