From repertoire to project: MediaWiki database and generative AI applied to railway right-of-way requalification projects.

Abstract

The article reports the making of the MediaWiki, a semantic network database that gathered project references of Architecture, Urbanism and Design and stood out as a creative work method guiding project solution for urban-landscape treatment of railway right-of-way in Brazil. In order to promote data refinement, we introduced Generative Artificial Intelligence to analyze and summarize a keywords preset list related to the cataloged projects in continuous improvement cycles. This iteratively and AI driven process gave feedback to generate a final result of increasingly targeted and coherent alternatives, finally incorporated at the “Catalog of Replicable Territorial Solutions”, which is aligned with ESG guidelines and UN-SDGs. In conclusion, we discuss the relevance of Generative AI as a facilitating tool to manage complex and large-scale database platforms as an educational methodology and research strategy. We also approach the challenges still present and potential methodological improvements for future applications in different territorial contexts and scales.