Urban Knowledge Graph

Urban knowledge graphs (UrbanKGs) integrate diverse urban data sources—including geographic information, points of interest, and textual descriptions—into a structured network to facilitate knowledge discovery and application in smart city initiatives. Current research emphasizes automated UrbanKG construction using large language models (LLMs) like GPT-4 and Llama, often incorporating techniques like instruction fine-tuning and graph convolutional networks to improve efficiency and accuracy. This work aims to overcome limitations of manual data integration and enable more sophisticated applications such as urban spatiotemporal prediction, disaster response (e.g., damage assessment), and urban planning. The resulting improved data accessibility and analytical capabilities are expected to significantly advance research and practical applications in urban informatics.