Final Presentation of the Teaching Project Data-driven Solutions for the Smart City Hamburg

12 February 2026, by Marten Borchers

Photo: Marten Borchers/UHH

On Wednesday, the final event of the teaching project Data-driven Solutions for the Smart City Hamburg took place. The interdisciplinary and transdisciplinary teaching format, which was held for the eighth time, is conducted jointly with Hamburger Hochbahn AG, the University of Hamburg, and the Hub of Computing & Data Science.

This teaching format provides students with the opportunity to gain practical experience, while also promoting skill development and innovation in various ways. For this purpose, collaboration took place with the Department of Mobility Strategy and Strategic Product Planning, represented by Aurelia Mennerich, the Sustainability Management Division, represented by Nienke Berger and Laureen Brieve, as well as the Department of Mobility Data Analytics & Modeling, represented by Jastine Bergelt, and the Department of Bus Operations Organization, represented by Mirco Kötzsch.

The group, consisting of Michelle Hallmann, Nina Strarovitova, Julia Wilhelm, Bahar Barikzei, and Patrick Leonhardt, in cooperation with Jastine Bergelt, Michael Thüne, and Mirco Kötzsch, examined how AI can support the analysis of route information to evaluate operational impacts during detours and station relocations. For this purpose, a digital dashboard, including a map, was developed that automates the parsing and analysis of text-based reports and quantifies impacts via queries to the station register and OpenStreetMap. The prototype simplifies impact analysis and systematizes reports.

Julien Graf, Erik Pauli, Sardorbek Sabirov, Marvin Schmidt, and Lukas Owen, together with Nienke Berger and Laureen Brieve, developed a solution for creating AI-supported risk profiles for product groups and items to identify and assess sustainability risks along the supply chain. For this, the existing process was examined, and a prototype was developed that, through connection to a search engine and a large language model, breaks down goods and products into their components and performs automated searches for human rights and environmental risks. The intelligent search supports the creation of risk profiles and prepares qualitative information and sources.

The group consisting of Yasmine Khamassi, Leon Sothmann, and Josef Hermann (+1) investigated with Aurelia Mennerich how an AI-based prototype can improve access to accessibility requirements for implementation managers and further users. For this purpose, a large language model was developed using the Retrieval-Augmented Generation architecture, which draws on various documents on accessibility requirements and makes this knowledge accessible via an intuitive chatbot interface.

Jan Krause, Heiko Bornholdt, Benjamin Klinkigt, Eva Bittner,  Martin Semmann and Marten Borchers are pleased with the successful completion and the results. A heartfelt thanks to everyone involved, and we look forward to further collaboration.