Multi‐Row Labeling With Semantic Analysis: A Case Study on Chinese <scp>POIs</scp>

ABSTRACT Labels are widely used in maps to convey verbal information for symbols and play a crucial role in aiding users' navigation and understanding of spatial context. Traditional labeling approaches mainly focus on ensuring label readability by placing them overlap‐free while maintaining visual coherence. However, these methods often fall short of Point of Interest (POI) labeling, particularly in the context of Chinese labels, due to the growing demand for detailed and informative maps with long or descriptive POI names. To address this challenge, it is essential to shorten label names and split long labels into multiple rows to achieve a more effective layout. In this work, we present a multi‐row label algorithm that introduces new quality constraints and incorporates linguistic and semantic analysis for Chinese label preprocessing, segmentation, and placement. We also demonstrate its application in a real‐world mapping platform, Meituan Map, which serves over 50 million monthly users. To inform the design of the labeling algorithm, we interviewed six domain experts and conducted a statistical analysis based on a dataset containing 160,297 POIs. This analysis confirms the necessity of multi‐row labeling and highlights practical challenges and considerations. Our results indicate that the label preprocessing phase of our algorithm can reduce the total character count by 41.15% and the word count by 43.24%. Comparative experiments show that our approach achieves superior label placement quality in terms of readability and visual clarity. However, this improvement may come at the cost of increased computational time. A user study involving 318 participants demonstrated that multi‐row methods or methods with label processing result in a user‐preferred label layout with comparable effectiveness on user tasks, although it may increase response time.