AIDA-Vis developed an AI-based interactive recommendation system that supports users in selecting suitable data visualizations. The project focused on how complex data, especially spatio-temporal datasets, can be presented in a way that enables domain experts to arrive more quickly at understandable, reliable, and individually tailored visualizations. To this end, the project combined approaches from active learning, preference-based learning, and reinforcement learning. The system was designed to infer user preferences from interactions and use them to generate recommendations for suitable charts and visualization parameters. The project idea was prototypically investigated in the context of the disy Cadenza data analytics platform and tested in various application scenarios. The goal was to reduce the manual effort required to create sophisticated visualizations, improve the transparency and comprehensibility of data-based analyses, and open up new possibilities for learning assistance functions in business and location intelligence systems.