Graph network evaluation metrics.

Historical urban spaces, imbued with profound historical and cultural significance, have evolved from ‘spaces’ into meaningful ‘places’. But they now face the risk of being eroded by rapid urbanisation and forgotten by today’s society, making it challenging to integrate them into modern life. UNESCO’s Historic Urban Landscapes (HUL) Recommendation highlights the importance of understanding the holistic, layered, and dynamic nature of urban heritage. Therefore, focusing on the ancient city district of Guangzhou as a case study, this research explores innovative approaches to integrating holistic semantics of urban places by merging contemporary semantics from geo-big data with historical and cultural semantics from documents and archives to create an urban knowledge model that bridges the gap between the present and the past. Furthermore, using knowledge graph embedding technology, we develop a model capable of entity prediction, similarity calculation, and query retrieval. We propose four key application scenarios for the implementation of the model. First, our research identifies potential cultural spatial connections that contribute to the joint preservation and promotion of historic urban places. Second, we develop a recommendation system that caters to users’ various requests, increasing the visibility of historical places. Third, we predict optimal locations for Time-Honored Brands. Finally, we identify visitor profiles to assist managers in meeting cultural promotion needs. To summarize, the integrated framework proposed in this study demonstrates both methodological efficacy and reusability. It not only helps to deeply explore the historical and cultural connotations, providing a scientific basis for urban planning and cultural inheritance, but also has the potential for enhancing the public’s awareness and participation in historical culture, promoting the sustainable development and prosperity of urban culture.