Scholarlink

<p>As global urbanization shifts from expansive growth to focused regeneration, assessing how various regeneration models impact urban vitality becomes crucial. Historic districts, marked by deep-rooted identities and rising development pressure, demand nuanced regeneration strategies that align preservation with modernization, and serve as testing grounds for these contrasting approaches. However, the comparative efficacy of government-led and market-driven approaches remains underexplored. This study assesses how divergent regeneration models shape urban vitality, using Weibo_Expressed Sentiment (WESI) and Weibo_Check-in density (WCDI) as key indicators. Focusing on Suzhou’s Jianjin Qiao Alley (government-led) and Shiquan Street (market-driven), the research evaluates the spatial-temporal impacts of regeneration. Its mixed-methods framework uses a quasi-experimental Difference-in-Differences (DID) design for robust causal identification, complemented by the machine learning-based Extreme Gradient Boosting (XGBoost) model to handle non-linear prediction and feature analysis. The study draws on geotagged social media check-ins and Points of Interest (POI) data from 2020 to 2024. It quantifies how built-environment elements influence regeneration performance and sentiment expression. Findings reveal a distinct trade-off: (1) the Market-Driven model was superior for improving public perception, causing a significant 0.029% increase in the WESI. (2) In contrast, the Government-Led model excelled at drawing public presence, driving a 0.303% increase in the WCDI, an impact of a much larger magnitude. (3) The predictive XGBoost analysis uncovers a non-monotonic effect where WESI peaks when the catering density index (PCDI) is in the 0.5 to 1.5 range, but turns negative after its value surpasses a threshold of 2. This study challenges conventional regeneration paradigms, uncovering temporal trade-offs between market efficiency and cultural sustainability. By introducing an integrated DID-XGBoost assessment framework, it quantifies the externalities of historic district regeneration, providing a diagnostic tool for optimizing heritage-compatible development.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Regeneration Efficiency Assessment and Predictive Comparison of Government-Led and Market-Driven Models in Historic Districts Via DID and XGBoost

  • 期刊论文

  • 发表日期：

    2025年11月11日

查看全文
  • Hong Ni,
  • Jinliu Chen,
  • Pengcheng Li

摘要

As global urbanization shifts from expansive growth to focused regeneration, assessing how various regeneration models impact urban vitality becomes crucial. Historic districts, marked by deep-rooted identities and rising development pressure, demand nuanced regeneration strategies that align preservation with modernization, and serve as testing grounds for these contrasting approaches. However, the comparative efficacy of government-led and market-driven approaches remains underexplored. This study assesses how divergent regeneration models shape urban vitality, using Weibo_Expressed Sentiment (WESI) and Weibo_Check-in density (WCDI) as key indicators. Focusing on Suzhou’s Jianjin Qiao Alley (government-led) and Shiquan Street (market-driven), the research evaluates the spatial-temporal impacts of regeneration. Its mixed-methods framework uses a quasi-experimental Difference-in-Differences (DID) design for robust causal identification, complemented by the machine learning-based Extreme Gradient Boosting (XGBoost) model to handle non-linear prediction and feature analysis. The study draws on geotagged social media check-ins and Points of Interest (POI) data from 2020 to 2024. It quantifies how built-environment elements influence regeneration performance and sentiment expression. Findings reveal a distinct trade-off: (1) the Market-Driven model was superior for improving public perception, causing a significant 0.029% increase in the WESI. (2) In contrast, the Government-Led model excelled at drawing public presence, driving a 0.303% increase in the WCDI, an impact of a much larger magnitude. (3) The predictive XGBoost analysis uncovers a non-monotonic effect where WESI peaks when the catering density index (PCDI) is in the 0.5 to 1.5 range, but turns negative after its value surpasses a threshold of 2. This study challenges conventional regeneration paradigms, uncovering temporal trade-offs between market efficiency and cultural sustainability. By introducing an integrated DID-XGBoost assessment framework, it quantifies the externalities of historic district regeneration, providing a diagnostic tool for optimizing heritage-compatible development.