Using natural language processing to evaluate local conservation text: A study of 624 documents from 303 sites of the World Heritage Cities Programme

The preservation of Outstanding Universal Value (OUV) at World Heritage sites, particularly in urban environments, faces significant challenges due to irreversible damage and lack of adequate attention. Analyzing local conservation documents is critical for assessing awareness of and compliance with OUV standards. Traditional evaluation methods, however, are resource-intensive and subject to inefficiencies and errors, particularly when dealing with large volumes of text. To address these issues, our study employs natural language processing (NLP) techniques to enhance both the size of the sample and the accuracy of the data, thereby enabling a more comprehensive analysis of conservation texts. We examined 624 documents from 303 pivotal sites in the World Heritage Cities Programme, analyzing multiple dimensions and sub-labels related to OUV. Our findings reveal distinct regional variations in OUV-related concerns, influenced by factors such as OUV pillars, criteria, document types, and revision frequencies. Additionally, our research highlights how disparities in wealth, size, industrial structure, and levels of scientific and educational development across different urban heritage contexts contribute to variations in conservation quality. This study provides an efficient and thorough methodology for reviewing local-level plans, which enhances the monitoring and protection of OUV.