A new method is proposed using area feature-tourism districts as analysis unit. Firstly, spatial-temporal behaviors of individual tourists are extracted from social media data. Secondly, the city’s tourism districts are extracted based on spatial-temporal behaviors. Finally, tourism districts are analyzed by 3 kinds of features — ourist activity features, tourist origin features and the structure features of tourism district network using tourism districts as nodes and tourist flow as edges. In the empirical research of Suzhou, 7 tourism districts are extracted based on spatial-temporal behaviors. The spatial structure of tourism districts is generally the same as “1-core-1-corridor-3-district” pattern in Suzhou tourism planning. The feature analysis of Suzhou tourism districts indicates that the Ancient City Tourism District and the Ancient Town Tourism District are the core tourism districts, which attracts tourists from various and distant origins. The tourism districts in and near Suzhou urban area attract more tourists. Suzhou tourism districts have already formed into a multi-core structure. This research shows the effectiveness of extracting city’s tourism districts based on social media data and researching city’s tourism with tourism districts as analysis unit, providing a new approach for research on urban tourism.
