Xinyu Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xinyu Wang has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Ying Long
School of Architecture and Hang Lung Center for Real Estate, Key Laboratory of Ecological Planning & Green Building, Ministry of Education, Tsinghua University, Beijing, China
2 shared publications Hui Wang
Department of Landscape Planning and Design, School of Art Design and Media, East China University of Science and Technology, Shanghai, China
1 shared publication- Tangqi Tu
Tsinghua University
1 shared publication - Wenzhu Li
School of Art Design and Media, East China University of Science and Technology, Shanghai, China
1 shared publication Xiangfeng Meng
School of Architecture, Tsinghua University, Beijing, China
1 shared publicationYing Long
School of Architecture and Hang Lung Center for Real Estate, Key Laboratory of Eco Planning & Green Building, Ministry of Education, Tsinghua University, Beijing, China
1 shared publication
Latest Publications
- 8 July 2026
- Published by SAGE Publications in Transactions in Urban Data, Science, and Technology
- Vol. 5 (3) , 280-302
Large language models for evidence-based planning: Evaluating an SLR-RAG framework for knowledge synthesis of urban vacant land
Urban challenges necessitate robust evidence synthesis, yet Large Language Models (LLMs) applied to urban studies face factual inconsistencies and a lack of domain-specific grounding. This paper proposes and evaluates a framework for RAG-enhanced LLMs by integrating up-to-date knowledge and policy-relevant evidence into model outputs. Urban Vacant Land (UVL) is selected as the case study, and a systematic literature review (SLR) is conducted to build a retrieval literature base. The results indicate that RAG significantly improves the accuracy of LLMs under structured automatic evaluation. However, although retrieval augmentation provides models with access to domain-specific evidence, its benefits for open-ended planning questions are not consistently reflected in expert ratings, particularly for consistency and creativity. The limited improvement can be attributed to the characteristics of the planning questions, the mismatch between textual information and the spatial data required for urban planning in current RAG pipelines, and potentially ineffective prompting that fails to elicit deeper reasoning. This study contributes to the field by elucidating the capabilities and limitations of LLMs and RAG in urban studies, highlighting that while text-only retrieval is insufficient for context-rich analysis, future advancements in spatially aware hybrid retrieval offer a promising pathway forward.
- 8 July 2026
- Published by SAGE Publications in Transactions in Urban Data, Science, and Technology
- Vol. 5 (3) , 280-302
Large language models for evidence-based planning: Evaluating an SLR-RAG framework for knowledge synthesis of urban vacant land
Urban challenges necessitate robust evidence synthesis, yet Large Language Models (LLMs) applied to urban studies face factual inconsistencies and a lack of domain-specific grounding. This paper proposes and evaluates a framework for RAG-enhanced LLMs by integrating up-to-date knowledge and policy-relevant evidence into model outputs. Urban Vacant Land (UVL) is selected as the case study, and a systematic literature review (SLR) is conducted to build a retrieval literature base. The results indicate that RAG significantly improves the accuracy of LLMs under structured automatic evaluation. However, although retrieval augmentation provides models with access to domain-specific evidence, its benefits for open-ended planning questions are not consistently reflected in expert ratings, particularly for consistency and creativity. The limited improvement can be attributed to the characteristics of the planning questions, the mismatch between textual information and the spatial data required for urban planning in current RAG pipelines, and potentially ineffective prompting that fails to elicit deeper reasoning. This study contributes to the field by elucidating the capabilities and limitations of LLMs and RAG in urban studies, highlighting that while text-only retrieval is insufficient for context-rich analysis, future advancements in spatially aware hybrid retrieval offer a promising pathway forward.
- 17 November 2025
- Published by Elsevier in Habitat International
- Vol. 167
Urban vacant land identification and its distribution rules of China's urban system based on fast segment anything model
- 17 November 2025
- Published by Elsevier in Habitat International
- Vol. 167
Urban vacant land identification and its distribution rules of China's urban system based on fast segment anything model
- Open Access
- 10 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 18 (19391404) , 8523-8535
Automatic Identification of Potential Renewal Areas in Urban Residential Districts Using Remote Sensing Data and GeoAI
Urban renewal is crucial for fostering revitalization and sustainable development in cities. Accurate identification of renewal areas in urban residential districts is essential for implementing effective renewal strategies. However, existing studies struggle with the automatic large-scale spatial classification of renewal areas due to their inherent complexity. Moreover, complete demolition and reconstruction are not always conducive to long-term sustainability. This study proposes an automated framework to identify potential renewal areas in urban residential districts, using remote sensing data and GeoAI, with Shanghai as a case study. The framework began by establishing classification rules to identify different renewal modes: retention, renewal, and demolition areas. The data labeling process was performed using the Segment Anything Model, followed by a comprehensive identification of different renewal modes through deep learning models, specifically DeepLabv3+. The results showed that (1) a total of 327.18 km² of retention areas, 130.76 km² of renewal areas, and 37.31 km² of demolition areas were automatically identified; (2) retention areas were evenly distributed across the city, while potential renewal areas were concentrated in the central urban districts, and demolition areas were primarily located in the suburban regions. This automatic identification framework significantly enriches the understanding of spatial complexity in urban systems and broadens the application of remote sensing technology in urban studies. The findings provide valuable insights for urban planners, highlighting priority regions for targeted renewal efforts.
- 10 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 18 (19391404) , 8523-8535
Open AccessAutomatic Identification of Potential Renewal Areas in Urban Residential Districts Using Remote Sensing Data and GeoAI
Urban renewal is crucial for fostering revitalization and sustainable development in cities. Accurate identification of renewal areas in urban residential districts is essential for implementing effective renewal strategies. However, existing studies struggle with the automatic large-scale spatial classification of renewal areas due to their inherent complexity. Moreover, complete demolition and reconstruction are not always conducive to long-term sustainability. This study proposes an automated framework to identify potential renewal areas in urban residential districts, using remote sensing data and GeoAI, with Shanghai as a case study. The framework began by establishing classification rules to identify different renewal modes: retention, renewal, and demolition areas. The data labeling process was performed using the Segment Anything Model, followed by a comprehensive identification of different renewal modes through deep learning models, specifically DeepLabv3+. The results showed that (1) a total of 327.18 km² of retention areas, 130.76 km² of renewal areas, and 37.31 km² of demolition areas were automatically identified; (2) retention areas were evenly distributed across the city, while potential renewal areas were concentrated in the central urban districts, and demolition areas were primarily located in the suburban regions. This automatic identification framework significantly enriches the understanding of spatial complexity in urban systems and broadens the application of remote sensing technology in urban studies. The findings provide valuable insights for urban planners, highlighting priority regions for targeted renewal efforts.
- 27 June 2023
- Published by SAGE Publications in Environment and Planning B: Urban Analytics and City Science
- Vol. 50 (7) , 1994-1997
Future shrinking cities on the globe: A projection map for 2020–2100 based on global gridded population dataset
Shrinking cities have become increasingly prevalent worldwide due to various factors, which pose serious challenges to affected areas in terms of population decline, economic decline, and spatial deterioration. While existing research studies have focused on identifying shrinking cities, there is a need for global projections to mitigate uncertainties in their growth trajectories. Spatially explicit population grids offer a new approach to identifying potentially shrinking cities with sufficient spatial resolution. By utilizing a global gridded population dataset from 2020 to 2100 under the SSP2 (Middle of the Road) scenario, we produce a global projection map for future shrinking cities. Among the total 19,024 natural cities, 9682 cities (50.9%) will face population decline and 1751 cities (9.2%) may lose more than half population by 2100. Cities in East Asia and East Europe may face serious population decline.
- 27 June 2023
- Published by SAGE Publications in Environment and Planning B: Urban Analytics and City Science
- Vol. 50 (7) , 1994-1997
Future shrinking cities on the globe: A projection map for 2020–2100 based on global gridded population dataset
Shrinking cities have become increasingly prevalent worldwide due to various factors, which pose serious challenges to affected areas in terms of population decline, economic decline, and spatial deterioration. While existing research studies have focused on identifying shrinking cities, there is a need for global projections to mitigate uncertainties in their growth trajectories. Spatially explicit population grids offer a new approach to identifying potentially shrinking cities with sufficient spatial resolution. By utilizing a global gridded population dataset from 2020 to 2100 under the SSP2 (Middle of the Road) scenario, we produce a global projection map for future shrinking cities. Among the total 19,024 natural cities, 9682 cities (50.9%) will face population decline and 1751 cities (9.2%) may lose more than half population by 2100. Cities in East Asia and East Europe may face serious population decline.
- Top Cited PapersOpen Access
- 12 September 2022
- Published by Springer Nature in Scientific Data
- Vol. 9 (1) , 1-13
Projecting 1 km-grid population distributions from 2020 to 2100 globally under shared socioeconomic pathways
Spatially explicit population grid can play an important role in climate change, resource management, sustainable development and other fields. Several gridded datasets already exist, but global data, especially high-resolution data on future populations are largely lacking. Based on the WorldPop dataset, we present a global gridded population dataset covering 248 countries or areas at 30 arc-seconds (approximately 1 km) spatial resolution with 5-year intervals for the period 2020–2100 by implementing Random Forest (RF) algorithm. Our dataset is quantitatively consistent with the Shared Socioeconomic Pathways’ (SSPs) national population. The spatially explicit population dataset we predicted in this research is validated by comparing it with the WorldPop dataset both at the sub-national and grid level. 3569 provinces (almost all provinces on the globe) and more than 480 thousand grids are taken into verification, and the results show that our dataset can serve as an input for predictive research in various fields.
- 12 September 2022
- Published by Springer Nature in Scientific Data
- Vol. 9 (1) , 1-13
Top Cited PapersOpen AccessProjecting 1 km-grid population distributions from 2020 to 2100 globally under shared socioeconomic pathways
Spatially explicit population grid can play an important role in climate change, resource management, sustainable development and other fields. Several gridded datasets already exist, but global data, especially high-resolution data on future populations are largely lacking. Based on the WorldPop dataset, we present a global gridded population dataset covering 248 countries or areas at 30 arc-seconds (approximately 1 km) spatial resolution with 5-year intervals for the period 2020–2100 by implementing Random Forest (RF) algorithm. Our dataset is quantitatively consistent with the Shared Socioeconomic Pathways’ (SSPs) national population. The spatially explicit population dataset we predicted in this research is validated by comparing it with the WorldPop dataset both at the sub-national and grid level. 3569 provinces (almost all provinces on the globe) and more than 480 thousand grids are taken into verification, and the results show that our dataset can serve as an input for predictive research in various fields.