Many important questions – about climate, health, or cities – cannot be answered by simply retrieving existing information. They require combining and transforming data in meaningful ways. At the same time, there are many possible data sources and many alternative geo-analytical paths to an answer.

GeoTrAnsQData develops a new way of answering geographic questions with data. Instead of searching for existing maps or manually figuring out how to analyse them, we enable users to generate answers by transforming available geodata sources – in ways that are explainable with respect to the question and justifiable in terms of the data used.

This project introduces transformative Geographic Question Answering (GeoQA): a novel approach that interprets questions as requests for analysis and suggests workflows to generate valid answers from existing geodata repositories.

By combining symbolic and sub-symbolic AI, knowledge graphs, and theories of geographic information, GeoTrAnsQData aims to make geospatial analysis not only more accessible, but also more transparent, opening up new possibilities for researchers, policymakers, and society.