Recent advances in large language models (LLMs) have expanded natural language reasoning and multimodal understanding but remain limited in grounding with 3D spatial environments. This project addresses that gap by developing a system that enables natural language interaction with indoor spatial data derived from light detection and ranging (LiDAR) point clouds and panoramic imagery provided by the client: ScanPlan. The system processes spatial data through a pipeline that includes room segmentation, geometric analysis, and object clustering. A structured query language lite (SQLite) database stores the structured information, which an AI agent queries using a reasoning framework that translates natural language into actionable commands. The system supports multimodal input, allowing users to interact via text or by selecting objects in 2D panoramas, which are then mapped to 3D point clouds using segment anything model 2 (SAM2). The interface combines a chat function with 2D and 3D viewers, making spatial data accessible to non-experts. While the prototype successfully answers a range of spatial and semantic queries, challenges remain in scaling room segmentation and handling complex multi-room relationships. The project demonstrates a step towards making rich 3Dbuilding data queryable through intuitive, language-based interaction.

With the Urban Heat Island Effect on the rise and 66% of the global population expected to live in the cities by 2050, new ways of building for this urban growth need to be researched. In this report the cooling effects of green and water structures on urban landscapes are examined. To realize this, we opted to recreate a 3D model of the chosen research area based on different
datasets. This can then be used in the simulation software ENVI-met to calculate the effect that these structures have on the microclimate. Firstly, we successfully made a 2D inventory based on the different components necessary to run the simulation.
Secondly, when converting from the 2D data to 3D data, errors occurred as regards to the tree and building datasets. Thirdly, the simulation failed due to a lack of interoperability between the datasets and softwares. Therefore, we are unable to answer the research question. However, after an in-depth assessment of the FAIRness of data, we conclude that by testing, failing and improving the
interoperability of datasets and softwares, we can change our perception of the data we collect and find new ways to store and visualize them.