Cultural heritage, especially those inscribed on the UNESCO World Heritage List, is meant to be valued by mankind and protected for future generations. Triggered by radical and sometimes disastrous Heritage-Related Events (HREs), communities around the world are actively involved on social media to share their opinions and emotional attachments. This paper presents exploratory data analyses on a dataset collected from Twitter concerning HREs in World Heritage that triggered global concerns, with cases of the Notre Dame Paris fire and the Venice flood, both in 2019. The spatiotemporal patterns of tweeting behaviours of online communities before, during, and after the event demonstrate a clear distinction of activation levels caused by the HREs. The dominant emotions and topics of people during the online debate are detected and visualized with pre-trained deep-learning models and unsupervised clustering algorithms. Clear spatiotemporal dynamics can be observed from the data collected in both case studies, while each case also demonstrated its specific characteristics due to the different severity. The methodological framework proposed and the analytical outcomes obtained in this paper could be used both in urban studies to mine the public opinions about HREs and other urban events for reducing risks, and by the Geo-AI community to test spatiotemporal clustering algorithms.

Mapping cultural significance of heritage properties in urban environment from the perspective of the public has become an increasingly relevant process, as highlighted by the 2011 UNESCO Recommendation on the Historic Urban Landscape (HUL). With the ubiquitous use of social media and the prosperous developments in machine and deep learning, it has become feasible to collect and process massive amounts of information produced by online communities about their perceptions of heritage as social constructs. Moreover, such information is usually inter-connected and embedded within specific socioeconomic and spatiotemporal contexts. This paper presents a methodological workflow for using semi-supervised learning with graph neural networks (GNN) to classify, summarize, and map cultural significance categories based on user-generated content on social media. Several GNN models were trained as an ensemble to incorporate the multi-modal (visual and textual) features and the contextual (temporal, spatial, and social) connections of social media data in an attributed multi-graph structure. The classification results with different models were aligned and evaluated with the prediction confidence and agreement. Furthermore, message diffusion methods on graphs were proposed to aggregate the post labels onto their adjacent spatial nodes, which helps to map the cultural significance categories in their geographical contexts. The workflow is tested on data gathered from Venice as a case study, demonstrating the generation of social perception maps for this UNESCO World Heritage property. This research framework could also be applied in other cities worldwide, contributing to more socially inclusive heritage management processes. Furthermore, the proposed methodology holds the potential of diffusing any human-generated location-based information onto spatial networks and temporal timelines, which could be beneficial for measuring the safety, vitality, and/or popularity of urban spaces.