Realistic characteristics and driving mechanisms of pseudo-human settlements in Chinese cities

This research investigates pseudo-human settlements (PHS) in Chinese cities, defined as information-based settlements constructed by residents using media. The study acknowledges the limitations of direct experience in understanding the entirety of the outside world, emphasizing the role of media in shaping perceptions. The concept of PHS builds upon Lippmann's 'pseudo-environment,' but extends it to encompass online human settlement activities characterized by authenticity, deviation, and informatization. The rapid urbanization in China provides a valuable context for this study, aiming to understand the subjective needs and preferences of residents regarding their living environment. The study leverages the Baidu index, a popular Chinese search engine, to collect data reflecting residents' online activities and preferences, supplementing the limitations of social media data. Existing research has explored PHS in specific contexts, but this study aims to provide a comprehensive analysis of its driving mechanisms at the national scale in China, building upon previous work that highlighted the positive correlation between pseudo-human settlements and real human settlement development levels, primarily driven by socioeconomic factors.

Existing literature demonstrates the use of social media data to understand urban space and residents' behaviors and emotions. Studies have used geo-tagged data from platforms like Weibo, Flickr, and Facebook to explore the relationship between urban space and residents' emotional responses and activities, as well as to identify the function and structure of real urban space. However, the subjectivity of social media data, potential information pollution from rumors and fake news, and the focus on specific application names as keywords, limit its comprehensive and objective description of PHS. This study addresses these limitations by utilizing the Baidu index, which offers a broader perspective by providing geographic data tags and capturing spontaneous search information, enhancing the objectivity of the index system by using functions as keywords.

The study selected 286 Chinese cities (municipalities, sub-provincial cities, and prefecture-level cities) for analysis. The Pseudo-Human Settlements Index (PHSI) was calculated using the Baidu index, aggregating search indices for common PHS classifications (home life, socialization, shopping, travel, entertainment, etc.) from January 1, 2015, to December 31, 2020. The index system was divided into ten functional structure indices, each with three related keywords, standardized using the Min-Max method. The PHSI was calculated by summing these standardized indices (Equation 1) and the proportion of each index in the PHSI was determined (Equation 3). The study also gathered data on realistic attributes of the cities, including statistical data (population, per capita GDP, industrial structure), nighttime light index, and air quality index from multiple sources. Spatial autocorrelation analysis (Global Moran's I and LISA) was used to analyze the spatial patterns of PHSI, while K-means cluster analysis grouped cities based on reality attributes and PHS functional structures. Geographic detectors were applied to examine the driving mechanisms of reality attributes on PHS development level and functional structure differentiation. This involved using factor detectors to determine the explanatory power of individual factors and interaction detectors to explore the combined effects of multiple factors.

The study revealed that PHS development levels exhibit notable hierarchical and spatial differences, with higher levels concentrated in eastern coastal regions and lower levels in western inland regions. Global Moran's I index showed weak positive spatial autocorrelation. Local spatial autocorrelation (LISA) analysis identified 'diffusion' characteristics in central cities of national urban agglomerations, 'sinking' characteristics in peripheral cities, 'polarization' characteristics in provincial capitals of remote provinces, and 'contagious' characteristics in prefecture-level cities of remote provinces. Key drivers of PHS development level, identified using geographic detectors, were nighttime light index, population size, and per capita GRP. The interaction between population size and per capita GRP showed the highest explanatory power (0.682). K-means clustering identified six functional structure types of PHS, with learning and education consistently showing high proportions. The functional structure of PHS exhibited a dynamic change process characterized by an inverted 'V' shaped life cycle (emergence-development-maturity-decline-extinction). The polarization of learning and education functions intensified over time, particularly in cities with low PHS development levels. Analysis of the correlation between PHSI and the degree of polarization revealed a significant negative correlation in 2019-2020, suggesting that low PHSI cities experienced a rapid increase in the polarization of PHS functional structures.

The findings highlight the significant influence of socioeconomic factors on PHS development, confirming previous qualitative analyses. The spatial distribution patterns reflect existing regional development disparities in China, with national development policies leading to spatial-temporal imbalances. While urban agglomerations exhibit hierarchical diffusion, remote provinces show polarization. This spatial imbalance underscores the need for coordinated development strategies, including addressing institutional obstacles and mitigating the potential negative effects of concentrated development in major cities. The analysis of PHS functional structure underscores the dynamic and evolving nature of residents' preferences, influenced by multiple socioeconomic factors. The study's focus on subjective preferences adds a new dimension to human settlement studies, providing insights into resident-centric urban planning.

This study provides a comprehensive understanding of the characteristics and driving mechanisms of PHS in Chinese cities. It emphasizes the importance of considering residents' subjective preferences in urban planning and highlights the need for coordinated development strategies to address regional disparities. The PHSI offers a valuable tool for monitoring and evaluating human settlements, providing insights into evolving preferences and guiding policy adjustments. Future research could explore the impact of specific policies on PHS development and investigate the long-term trends of PHS functional structure changes.

The classification results of the K-means clustering method for PHS functional structure might vary depending on the study period due to the dynamic nature of urban PHS. Longer-term observations are needed to establish more stable results. The reliance on Baidu index data might not fully capture all aspects of online activities and could be influenced by search algorithm biases. Further investigation could incorporate additional data sources to enhance the robustness and generalizability of the findings.