| id |
caadria2025_352 |
| authors |
Lu, Tenglong and Tong, Ziyu |
| year |
2025 |
| title |
Graph-Based Typology Identification and Analysis of Urban Form: A case study of Nanjing main district |
| source |
Dagmar Reinhardt, Christiane M. Herr, Anastasia Globa, Jielin Chen, Taro ?Narahara, Nicolas Rogeau (eds.), ARCHITECTURAL INFORMATICS - Proceedings of the 30th CAADRIA Conference, Tokyo, 22-29 March 2025, Volume 4, pp. 39–48 |
| summary |
Rapid urbanization has led to the expansion and increased complexity of cities, requiring a deeper investigation into urban morphology. Traditional methods for analyzing urban typologies often treat plots or buildings as isolated entities, ignoring their relationships with the surrounding environment. This study aims to avoid this drawback by constructing a graph model that incorporates spatial relationships between urban elements, while also validating the rationality of using graph models in urban morphology research. The study constructs a graph model based on cities, using buildings as nodes and attaching related information as features. Edges are derived from the influence zones, symbolizing their adjacency relationships. Subsequently, a Graph Convolutional Network (GCN) is used to process the graph model to obtain low-dimensional feature representations, which are then input into clustering algorithms to produce classification results for different morphologies. We take Nanjing as a case study and also compares the results of the proposed method with those of non-graph clustering. The findings ultimately demonstrate that the clustering method based on graph models generates more coherent and meaningful classifications, proving the effectiveness of the method and providing a new perspective for understanding urban spatial characteristics. |
| keywords |
Graph Model, Graph Convolutional Networks, Typo-morphology, Urban Morphology, Clustering Algorithms |
| series |
CAADRIA |
| full text |
 file.pdf (2,035,555 bytes) |
| last changed |
2025/03/21 12:09 |
|
