AI-Powered Planning Platforms Reshape Historic Urban Renewal
In the evolving landscape of urban development, a quiet but profound transformation is taking place—not through bulldozers and steel frames, but through algorithms, sentiment analysis, and deep-learning models. A recent study published in Modern Information Technology demonstrates how next-generation information technologies, particularly artificial intelligence (AI) and big data analytics, are being strategically deployed to align historic urban renewal and characteristic town planning with public sentiment and cultural preservation.
The research, led by Tang Shaohua of Hunan Vocational College of Engineering, introduces an innovative planning assistant platform that leverages user-generated content from online travel and social media platforms to extract behavioral patterns and visual preferences. By analyzing textual reviews and geotagged images, the system translates subjective public opinion into quantifiable data, offering urban planners and municipal authorities a new lens through which to understand the emotional and aesthetic dimensions of historic sites.
This approach marks a significant departure from conventional urban planning paradigms, which have often prioritized economic efficiency or architectural vision over community engagement and cultural continuity. The consequences of such top-down strategies are visible in cities worldwide: historic districts stripped of their authentic character, heritage sites overwhelmed by generic commercialization, and public spaces that fail to resonate with local identity. Tang’s platform seeks to counteract these trends by embedding public perception directly into the planning workflow.
At the core of the system is a user perception data management module, designed to capture, classify, and interpret massive volumes of unstructured data. The process begins with web crawling—automated scripts harvest user comments and photos from popular platforms like Ctrip, focusing on specific landmarks such as Aiwan Pavilion in Changsha. These raw inputs are then processed through a dual AI pipeline: natural language processing (NLP) handles textual sentiment and keyword extraction, while computer vision models categorize visual elements into predefined landscape typologies—buildings, vegetation, water bodies, ornamental features, pathways, hardscapes, and rock formations.
Crucially, the platform does not merely count occurrences. It contextualizes them. For instance, in the case of Aiwan Pavilion—a culturally significant site nestled within Yuelu Mountain—the analysis revealed that while physical structures like pavilions and traditional towers drew consistent attention, it was natural elements—particularly maple trees and mountainous backdrops—that dominated public sentiment. Keywords such as “maple,” “forest,” “natural park,” and “Yuelu Mountain” appeared with striking frequency, underscoring a public affinity for the interplay between built heritage and surrounding ecology.
This insight carries direct planning implications. Rather than focusing solely on restoring masonry or repainting eaves, planners might prioritize preserving sightlines to the mountain, enhancing autumn foliage displays, or designing pedestrian routes that frame key natural vistas. The system thus shifts the planning focus from static preservation to dynamic experience—recognizing that heritage is not just about what remains, but how it is perceived and remembered.
The methodology also introduces a new form of participatory governance. Traditionally, public input in urban planning has been limited to formal consultations, surveys, or town hall meetings—processes that often exclude large segments of the population, particularly younger or digitally active demographics. By tapping into existing online behaviors, Tang’s platform captures organic, real-time feedback from individuals who may never attend a municipal hearing but regularly share their experiences online. In effect, it democratizes data collection, turning everyday tourists and residents into inadvertent co-designers of the urban fabric.
Importantly, the platform adheres to ethical data practices. The study emphasizes that all collected data is publicly available and anonymized; no private user information is harvested. Moreover, the AI models are designed to interpret general trends rather than individual opinions, ensuring that the outputs serve collective insight without compromising personal privacy.
The integration of narrative environment theory further elevates the platform’s analytical depth. Drawing from disciplines like human geography and cultural studies, the concept of “narrative space” treats urban landscapes not as inert backdrops but as vessels of stories—layered with historical memory, personal experience, and social meaning. By mapping user comments onto spatial and thematic categories (e.g., “disappeared material elements,” “non-material landscape features,” “human activities”), the system begins to reconstruct the narrative skeleton of a place. This allows planners to identify which stories are being told, which are fading, and which need reinforcement.
For example, the near-zero mentions of “vanished elements” at Aiwan Pavilion suggest a strong continuity between current conditions and public memory—indicating successful preservation. Conversely, a surge in nostalgic references to “what used to be there” at another site might signal a need for interpretive interventions, such as augmented reality overlays or commemorative signage, to bridge memory gaps.
Technologically, the platform’s reliance on third-party AI APIs—specifically Baidu AI Cloud for image classification and NLP—ensures robustness without requiring in-house model development. This modular architecture enhances scalability; the same framework can be applied to any historic district with sufficient online data. Future iterations could incorporate multimodal fusion techniques, combining text, image, and even geospatial metadata to generate richer perceptual maps.
From a policy standpoint, the implications are far-reaching. Municipal governments face increasing pressure to balance development with heritage conservation, often under tight budgets and political constraints. Tools like this offer evidence-based justification for preservation decisions, countering purely economic arguments with quantified public sentiment. They also enable rapid feedback loops: after a renovation project is completed, planners can re-run the analysis to assess whether public perception has improved, remained stable, or deteriorated—creating a cycle of iterative, responsive design.
Critically, the research avoids the techno-utopian trap. Tang acknowledges limitations: online reviewers may not represent the full demographic spectrum (e.g., elderly residents or low-income groups with limited digital access), and AI models can misclassify culturally specific elements. The platform is positioned not as a replacement for human judgment but as a decision-support tool—one that augments professional expertise with empirical public insight.
This measured, human-centered application of AI aligns with broader shifts in smart city discourse. Where early smart city initiatives often fixated on sensor networks and real-time traffic optimization, contemporary approaches increasingly emphasize equity, identity, and cultural sustainability. Tang’s work exemplifies this maturation, demonstrating that the most impactful urban technologies are those that listen—truly listen—to the voices of the people who inhabit the spaces being shaped.
Looking ahead, the methodology could extend beyond tourism sites to residential historic districts, where emotional attachment is even more profound. Imagine applying similar analysis to neighborhood forums, local review sites, or even municipal complaint systems to understand how residents feel about proposed changes to street layouts, façade regulations, or public amenities. The potential for conflict mitigation alone is significant: if planners can anticipate emotional resistance to a design before construction begins, they can adjust plans proactively, saving time, money, and community trust.
Moreover, as generative AI advances, future versions of the platform might simulate public reactions to proposed designs before they are built—generating predictive sentiment maps based on visual renderings and textual descriptions. While such capabilities remain speculative, the foundational work presented here provides the data infrastructure and conceptual framework to make them viable.
In an era where urban identity is increasingly fragile—threatened by homogenizing global development patterns and climate-driven displacement—tools that safeguard intangible heritage are as vital as those that protect physical structures. Tang Shaohua’s research offers a compelling blueprint for how technology, when grounded in cultural sensitivity and public participation, can serve not just efficiency, but meaning.
As cities worldwide grapple with the dual challenges of modernization and memory, this integration of AI, big data, and narrative theory may well become a standard practice in the planner’s toolkit—not as a flashy innovation, but as a necessary compass for navigating the complex emotional terrain of place.
Author: Tang Shaohua
Affiliation: Hunan Vocational College of Engineering, Changsha 410151, China
Journal: Modern Information Technology
DOI: 10.19850/j.cnki.2096-4706.2021.03.028