The AI4Copernicus health bootstrapping services tackle existing challenges in public health and air quality by leveraging Earth observation and on-site measurement data. The services primarily center around the probabilistic downscaling of air quality and atmospheric composition (AC) model results. This downscaling process enables the generation of hyper-local short to medium-term air pollution forecasts, offering street-level accuracy in densely populated regions. Today, air quality information is usually shared through weather apps.
Recent developments in chatbot technologies underline the potential of language as an interface to technology. In the AQQA project we aimed to test this. In the scope of AQQA the use of the linked geospatial data tools was investigated to prepare air quality data for language-based interaction. Furthermore, the integration of the health-related bootstrapping service into AQQA to access air quality data of a higher spatial resolution was evaluated. The overall service architecture made use of a knowledge graph and is using GeoTriples for RDF conversion, the related RDF store Strabon and Sextant for visualisation and interaction. For language- based queries, the OpenAI API was applied.