Next Article in Journal
Climate Responsive Green Building Strategies in Circular Cities: A Comparative Study for Two Regions
Previous Article in Journal
Farmers’ Adoption of Water Management Practice for Methane Reduction in Rice Paddies: A Spatial Analysis in Shiga, Japan
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 17
Issue 8
10.3390/su17083465
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Exploring the Mental Health Benefits of Urban Green Spaces Through Social Media Big Data: A Case Study of the Changsha–Zhuzhou–Xiangtan Urban Agglomeration

by
Zhijian Li
and
Tian Dong
*
School of Architecture and Art, Central South University, Changsha 410075, China
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(8), 3465; https://doi.org/10.3390/su17083465
Submission received: 16 February 2025 / Revised: 8 April 2025 / Accepted: 11 April 2025 / Published: 13 April 2025
(This article belongs to the Topic Sustainable Built Environment, 2nd Volume)
Browse Figures
Versions Notes

Abstract

Urban green spaces (UGSs) provide recreational and cultural services to urban residents and play an important role in mental health. This study uses big data mining techniques to analyze 62 urban parks in the Changsha–Zhuzhou–Xiangtan urban agglomeration (CZXUA) based on data such as points of interest (POIs), areas of interest (AOIs), and user comments from the popular social media platform Dianping. In addition, the authors apply sentiment analysis using perceptual dictionaries combined with geographic information data to identify text emotions. A structural equation model (SEM) was constructed in IBM SPSS AMOS 24.0 software to investigate the relationship between five external features, five types of cultural services, nine landscape elements, four environmental factors, and tourist emotions. The results show that UGS external features, cultural services, landscape elements, and environmental factors all have positive effects on residents’ emotions, with landscape elements having the greatest impact. The other factors show similar effects on residents’ moods. In various UGSs, natural elements such as vegetation and water tend to evoke positive emotions in residents, while artificial elements such as roads, squares, and buildings elicit more varied emotional responses. This research provides science-based support for the design and management of urban parks.

Graphical Abstract

1. Introduction

Rapid urbanization, accompanied by a sharp increase in urban populations, has exacerbated environmental challenges in cities, including air pollution, ecological degradation, overcrowding, and deteriorating living conditions [1]. The absence of human-centered urban development has had a severe negative impact on residents’ mental health [2], and various psychological issues have emerged as significant public health concerns that affect economic development and social stability [3].
UGS, as a crucial component of green infrastructure, has had a positive impact on residents’ mental health [4]. The multi-sensory stimulation provided by UGS promotes physical and mental recovery [5]. Natural parks help alleviate stress, improve mood [6], and foster mental restoration [7]. Higher vegetation coverage in UGS has been associated with lower depression rates [8]. UGS characteristics—such as structure, biodiversity, and naturalness—not only mitigate urban environmental issues but also relieve emotional stress and improve residents’ mental health by encouraging physical activity and enhancing social interactions [9]. Therefore, understanding their experiences and perceptions of green spaces is crucial for urban planners and developers.
Recent research on the impact of UGS on residents’ mental health has focused on specific characteristics, aspects, and functions of green spaces. For instance, Jaeyoung Ha et al. employed mental health data from over 6000 individuals collected by the Chicago Department of Public Health, finding that the dispersed distribution of many small green spaces provided greater mental health benefits to surrounding residents than a few large green spaces [10]. Brazilian scholar Adriano Bressane statistically analyzed online survey data, finding that multiple dimensions of natural contact in UGS (e.g., naturalness, intensity, duration, frequency, and infrastructure) were linked to mental health [11]. In China, Ran Zhang developed an evaluation system based on three aspects, aesthetic features (AFs), recreational facilities (FRs), and convenience facilities (FCs), to explore strategies for enhancing park quality to promote mental health and equity [12]. However, these studies often lack a comprehensive exploration of the various elements within green spaces that influence residents’ mental health. Additionally, traditional data collection methods—such as questionnaires and telephone surveys—have limitations, including restricted sample populations, low efficiency during data collection, small sample sizes, limited real-time capabilities, and high subjectivity.
With advances in satellite remote sensing technology and widespread internet access, foundational data on UGS can now be easily obtained from digital maps [13]. In addition, social media reviews offer first-person, geolocated, and multi-dimensional descriptions of landscape features, user experiences, service quality, and cost-effectiveness, as well as visitors’ emotions and the timing of their posts [14]. Analysis of user reviews through lexicon construction enables the assessment of the frequency of perceptions and emotional tendencies toward different UGS elements [15]. The large volume of social media reviews also provides a rich data source, enhancing the comprehensiveness, representativeness, and real-time accuracy of research on UGS.
This study aims to use geographical information and social media data to analyze residents’ perceptions and emotional experiences in UGS. It establishes a methodology to examine the impact of various UGS elements on mental health. With a focus on typical UGS in the CZXUA in central China, the study uses geographic coordinates of the parks and social media reviews from the platform Dianping as primary data sources. By constructing a Chinese lexicon and sentiment analysis library, the research identifies different types of perceptions and emotions expressed by residents. SEM is then employed to analyze the influence of UGS elements on residents’ mental health.
The CZXUA covers an area of 28,200 km2 (Figure 1). It is an essential part of the urban cluster in the middle reaches of the Yangtze River and comprises the cities of Changsha, Zhuzhou, and Xiangtan. The CZXUA plays a crucial role in regional development [16].
The CZXUA has a population of 17.07 million, with an urbanization rate of 76.97%, and in the first three-quarters of 2024, the regional GDP of CZXUA reached RMB 1600.83 billion, with Changsha accounting for 68.60% of the total. The UGS in the CZXUA covers 17,600 km2, with a 62.41% greening rate in built-up areas. The CZXUA includes China’s largest green heart area, which provided 528.32 km2 of green space (1.87% of the total area of the agglomeration) in 2024.
To ensure a sufficient data sample for each park, small UGSs lacking data in digital maps and parks with fewer than 100 reviews on Dianping.com were excluded. Urban parks with fewer than 100 online reviews were excluded to ensure methodological rigor. This threshold addressed data sparsity, as limited reviews hinder perception lexicon construction by failing to capture diverse visitor perceptions and amplifying outlier bias. SEM with 23 latent variables required a minimum sample size (5–10 times the free parameters) for stable parameter estimation, achieved by excluding parks with insufficient reviews. Prioritizing parks with more than 100 reviews improved the representativeness of our sample, as these parks are more likely to reflect mainstream visitor experiences and reduce potential biases from atypical or underutilized green spaces. Ultimately, the study selected the 62 UGSs with the highest number of reviews as the research areas. As of July 2024, the selected UGSs covered a total area of 193.91 km2, accounting for 1.10% of the total UGS in the CZXUA.

2. Literature Review

Analysis of Urban Residents’ Perceptions, Activities, and Emotions Based on Social Media Textual Analysis

Several scholars have analyzed the impact of urban green space elements on residents’ mental health, often based on traditional methods like surveys. For example, Delgado-Serrano, M. (2024) analyzed questionnaire feedback from 632 participants to study the effects of UGS on mental health and psychological stress in two medium-sized Spanish cities [17]. Wang, R.Y. (2023) conducted stratified random household sampling in 29 districts near the Connswater Community Greenway (CCG) in Belfast, UK, to investigate the links between the greenway and mental health [18]. Colleen E. Reid (2022) distributed questionnaires to communities in Denver, Colorado, during the COVID-19 pandemic, receiving 912 valid responses, to analyze perceptions of green space use, quality, and mental health relationships [19]. Zhang, J.G. (2022) carried out a three-month household survey in five districts of Nanjing, China, to study the biopsychosocial pathways connecting UGS with mental health [20]. Yigitcanlar et al. (2020) surveyed 300 residents in Tabriz, Iran, to explore the relationship between park visitation frequency and emotional well-being [21], while Yuen and Jenkins (2019) analyzed subjective well-being changes in 94 visitors before and after visiting three urban parks in Mountain Creek, Alabama [22]. However, these traditional methods lack efficiency, generalizability, and real-time accuracy.
In recent years, emerging big data technologies have increasingly been used in data collection for related studies [23]. Social media data have been widely applied in urban research [24]—from urban morphology and structure to human behavior—and are increasingly used in environmental studies [25,26]. For example, Walter, M. (2023) analyzed 140,000 Google Maps reviews of 285 Philadelphia parks from 2011 and 2022 to explore public perceptions of park quality and environmental justice [27]. Shi, L. (2024) collected 3006 reviews from Dianping.com to investigate factors affecting recreational experiences along the Hutuo River Corridor in Shijiazhuang, China [28]. Social media data contain geolocation information along with textual information on user activities, perceptions, and emotions [29]. These data provide large sample sizes with minimal bias and a human-centered focus. With the rapid development of big data and deep learning technologies, scholars are increasingly using machine learning and natural language processing (NLP) techniques [30] to conduct keyword analysis, quantify perceptions of cultural services, and analyze emotions in social media data [31,32].

3. Method and Data

In the analytical investigation, this study hypothesizes that the following four major categories of factors and their 23 sub-elements collectively influence residents’ psychological well-being (Figure 2):
(1)
The external characteristics of UGS affect residents’ emotions;
(2)
The types of cultural services provided by UGS affect residents’ emotions;
(3)
The types of landscape elements in UGS affect residents’ emotions;
(4)
The external environmental factors of UGS affect residents’ emotions.
This study follows four steps: data acquisition and preprocessing, lexicon construction, text analysis and sentiment analysis, and statistical analysis (Figure 3).
Data Acquisition and Preprocessing: Point of interest (POI), area of interest (AOI), and normalized difference vegetation index (NDVI) data for the selected parks were obtained from Baidu Maps and Google Earth Engine (GEE). Resident review text data was collected from Dianping.com and categorized accordingly.
Lexicon Construction: After segmenting the resident review texts using the “Jieba” word segmentation tool, a lexicon of UGS elements was constructed. This lexicon includes five categories for external characteristics, five for cultural ecosystem service elements, nine for landscape elements, and four for external environmental factors.
Text and Sentiment Analysis: Using Python® 3.1.15, words from each lexicon category were matched with segmented words in the review texts to conduct a text analysis. This process identified content related to the 23 categories of UGS elements. Sentiment scores were then assigned to the overall text and specific elements within each review using the SnowNLP library.
Statistical Analysis: Pearson correlation coefficients (PCCS) and SEM were used to analyze the data, resulting in relevant statistical charts and tables.

3.1. Data Acquisition and Preprocessing

The first part involves the collection of geographical information related to UGS. Based on a literature review, four categories of spatial characteristics relevant to residents’ mental health were identified (Table 1). Baidu Maps, a powerful online mapping service provider in China, offers a vast amount of POI and AOI data, covering information on dining, shopping, accommodation, attractions, and more to meet diverse user needs [33]. Using Baidu Maps’ API, this study extracted AOI data [34] for 62 UGSs, including park names, coordinates, and secondary classifications, updated as of June 2024. To ensure a structured analysis, the study classified green spaces into five types based on secondary AOI classifications using the categorization methods of Kong et al. and Şenik and Uzun [35,36] (Table 2). Additionally, NDVI data were collected to reflect vegetation coverage, indicating the level of greenery accessible to residents [37,38]. These data were sourced from GEE, a cloud computing platform developed by Google, Carnegie Mellon University, and the U.S. Geological Survey, which specializes in processing satellite images and other Earth observation data [39]. The NDVI data used in this study were sourced from GEE and calculated using Sentinel-2 satellite imagery (bands B8 and B4), providing the maximum NDVI value [40] for 2024 at a 30 m spatial resolution.
The second phase of data acquisition involves the collection of relevant data from social media platforms. Dianping.com, China’s leading platform for local life information and transaction, provides users with a wealth of merchant information, consumer reviews, and other user-generated content. Renowned for its authentic and trustworthy reviews and convenient transaction services, the platform allows users to freely share evaluations and experiences [47]. With a large user base, abundant data, and fast updates, Dianping.com effectively captures residents’ perceptions and opinions of UGSs [48]. Using the platform’s API, this study extracted 104,795 reviews from 62 UGSs, including fields such as park name, username, rating, review text, and review date [49]. After excluding empty, irrelevant, and duplicate reviews, a total of 103,901 reviews from July 2023 to July 2024 were selected for analysis.

3.2. Lexicon Construction

In analyzing the collected social media review texts, the first step is lexicon construction [50]. Unlike English, where words are separated by spaces, Chinese words are connected consecutively and must first undergo segmentation. For this purpose, the widely used open-source Chinese word segmentation tool “Jieba” (https://github.com/fxsjy/jieba, accessed on 28 August 2024) was used. Known for its efficiency, accuracy, and flexibility, Jieba is developed in Python and widely applied in NLP and text analysis to offer effective segmentation of Chinese text [51]. This study employed Jieba to segment each review [52].
To construct the lexicon framework, this study drew upon relevant research by Cao S. et al. [15,53,54,55,56,57], identifying one category of external characteristics, five types of cultural ecosystem service elements, nine types of landscape elements, and four types of external environmental factors (Table 3, Table 4, Table 5 and Table 6).
The third step is lexicon matching. High-frequency words from the segmented Jieba results were manually categorized into the 19 categories of the lexicon framework, with each category assigned 10–15 words. The Word2Vec algorithm was then applied to expand the word list [58]. In this study, Word2Vec identified words in the review texts that were semantically similar to those already listed in the lexicon, ranking them by frequency and adding them to the lexicon based on a similarity threshold of 80%. Finally, the newly added words were manually filtered to remove any irrelevant or semantically inconsistent terms, resulting in a finalized lexicon.

3.3. Text and Sentiment Analysis

The lexicon categories were matched with the segmented words from the review texts using Python [59]. Each review was analyzed to determine which categories of elements were present, providing information on the attributes referenced in each review. This result was then used to analyze the frequency of each element category. Additionally, sentiment analysis was performed on the textual content associated with each element category in each review [60], yielding sentiment scores for the elements mentioned.
For sentiment analysis, the SnowNLP library was used [61]. First, large amounts of textual data are labeled, followed by the extraction of text features, such as word frequency and word type. A Naive Bayes model is employed to train on these features, generating the probability of each feature with a specific sentiment category. Finally, these probabilities are aggregated to produce a sentiment score, classifying the text as either positive or negative [62,63]. Using this method, each review received an overall sentiment score and a sentiment score for the associated elements on a scale from 0 to 1. Scores between 0 and 0.5 were classified as negative sentiment, while scores between 0.5 and 1 were classified as positive sentiment. The higher the score was, the more positive the sentiment.

3.4. Statistical Analysis

The perception frequency of each element category was calculated by dividing the number of reviews mentioning a particular element in a park by the total number of reviews for that park [64]. Comparing the perception frequencies across different types of parks enabled insights into the structural characteristics of the ecosystem services in each park category.
To analyze the correlation between park geographic data, sentiment scores for each element (as obtained in Section 3.3.), and the overall sentiment scores of the review texts, IBM SPSS® Statistics was used [65]. The PCCS was applied to quantify the correlation strength between variables [66]. Typically, a correlation coefficient of 0–0.2 indicates no or very weak correlation, 0.2–0.4 weak correlation, 0.4–0.6 moderate correlation, 0.6–0.8 strong correlation, and 0.8–1 very strong correlation [67].
In statistical analysis, the p-value of the PCCS is used to test the significance of correlation. The smaller the p-value is, the more significant the correlation, with values below 0.05 generally considered significant and values below 0.01 regarded as highly significant [68].
To comprehensively assess the impact of various elements on residents’ mental health, this study employed SEM [69]. SEM integrates confirmatory factor analysis and path analysis, incorporating both measurement models and structural models. It surpasses traditional single-factor approaches by handling multiple dependent variables and estimating multi-factor structures and relationships, including those between observed and latent variables as well as between latent variables [70].

Structural Equation Modeling

SEM is a probabilistic model that integrates multiple predictor and response variables into a causal network. Typically represented by path diagrams, SEM indicates directional relationships between observed variables through arrows. These relationships are expressed through a series of structural equations that correspond to the paths in the model [71]. In existing research on the impact of UGS on mental health, SEM has been frequently used to study individual or specific elements, with fewer studies addressing the systematic and comprehensive effects of multiple elements. For example, Zhao, Y.W. (2024) used SEM to analyze questionnaire data, investigating the relationship between three types of parks and individual subjective well-being [72]. Similarly, Bev Wilson (2024) employed SEM to test the indirect effects of social cohesion and socioeconomic status on mental health [73]. This study’s use of SEM is innovative, as it not only capitalizes on the large sample size and rich information available in social media data but also incorporates multiple green space elements. This approach enables a comprehensive evaluation of the impact of park elements on residents’ mental health.

4. Results

4.1. Overview of Park Sentiment Distribution

Figure 4 illustrates the spatial distribution of the average sentiment scores for visitors to the parks studied. The heatmap shows sentiment hotspots primarily concentrated in the northern part of the CZXUA, around the urban areas of Changsha and between the three cities. This pattern indicates that parks in densely populated urban and adjacent areas are more likely to evoke positive emotions. In contrast, cold spots are located in the peripheral, less populated suburban areas, suggesting a weaker impact of these parks on mental health.
According to the method described earlier, the analysis of 103,901 reviews reveals that the number of reviews expressing positive emotions is approximately five times higher than those with negative sentiments. Comprehensive parks and recreational parks received the highest number of reviews, while cultural relics parks and ecological parks garnered less than one-third of the reviews received by the first two categories. Community parks had the fewest reviews (Figure 5).
When examining the sentiment scores by park category (Figure 6), sentiment scores for all parks range from 0 to 1, with an overall average of 0.83 and a standard deviation of 0.11. Reviews with sentiment scores in the ranges of 0–0.02 and 0.94–1 accounted for the largest proportions, with over 50% of visitors displaying high positive sentiment in their evaluations.

4.2. Distribution of Resident Sentiments Across Different Types of Parks

Based on sentiment analysis of review texts from the social media platform Dianping.com, the average sentiment score for residents in each park was calculated. The results were categorized according to the park classification in Table 2, with the distribution of average sentiment scores for each type of park shown in Figure 7.
As seen in Figure 7, most parks received positive average sentiment scores from residents, except for a small number of cultural relics parks. Comprehensive parks not only had the highest maximum sentiment scores but also a higher overall average sentiment score, with the lowest sentiment score still being the highest among all categories. This suggests that comprehensive parks are more likely to elicit positive emotions from most residents. Community parks ranked second, showing similarly high overall sentiment scores, making them another park type likely to enhance residents’ moods. Cultural relics parks, despite a high maximum sentiment score, had the lowest minimum score, with a broad range between the upper and lower quartiles. The sentiment distribution for cultural relics parks was skewed to the right, indicating a wide variability in emotional responses, likely influenced by external factors. Ecological and recreational parks exhibited similar sentiment score distributions, although ecological parks had more extreme high and low scores, indicating a greater susceptibility to other factors than recreational parks. Recreational parks, with a generally higher sentiment score, appear more likely to evoke positive emotional responses in residents than ecological parks.

4.3. External Characteristics of Green Spaces

In this section, the external characteristic of “transport conditions” from the lexicon is combined with the four geographical spatial characteristics of green spaces to describe the external characteristics of urban parks. Using IBM SPSS® Statistics, the PCCS between all the external characteristics of UGS and the corresponding average resident sentiment scores was calculated, as shown in Table 7.
From Table 7, it can be seen that all five external characteristics exhibit significant positive correlations with residents’ average sentiment scores, as indicated by p-values below 0.05. Transport conditions, availability, and boundary shape have relatively high PCCS values, indicating a moderate correlation with resident sentiment. The other characteristics, while still significant, show weaker correlations. These findings suggest that convenient transportation, access to green vegetation, and more irregular park boundary shapes contribute to a positive impact on residents’ emotions.

4.4. Ecosystem Cultural Service Characteristics of Different Types of Green Spaces

For the five types of UGS listed in Table 2, the perception frequencies of ecosystem cultural service elements were calculated for each park type, as shown in Figure 8.
Out of the 62 parks, 14 are comprehensive parks. For these, outdoor workouts have the highest frequency, closely followed by aesthetic appreciation. Social interaction and history/culture also exceed the average frequency, indicating that comprehensive parks deliver a variety of ecosystem cultural services. Most residents tend to engage in physical activities like walking, running, and sports, while the aesthetic layout of the landscape design is highly valued, making these parks important locations for both social and cultural activities.
Cultural relics parks, numbering 11, display the highest frequency for history and culture, while other elements fall below average. This shows that historical and cultural elements are the most important ecosystem cultural services in cultural relics parks, as they are the main attractions for most visitors.
Ecological parks, totaling 21, show a high frequency of aesthetic appreciation, followed by outdoor workouts, while social interaction also reaches the average frequency. This indicates that the aesthetic layout of ecological parks is highly attractive to residents. Ecological parks emphasize the preservation of the natural environment, aiming to harmonize interactions between people and nature, which is a key reason residents visit these parks. Additionally, many residents enjoy engaging in outdoor exercise and social activities in these natural settings.
Thirteen parks are categorized as recreational parks. In these parks, recreational activities have the highest frequency, while social interaction approaches the average frequency. This aligns with the recreational theme of these parks, as entertainment activities are the primary service provided. Many residents also consider these parks to be good places for socializing.
Community parks, of which there are three, exhibit an overall frequency distribution closest to the average, indicating that they provide a balanced range of ecosystem cultural services across various elements. Outdoor workouts, recreational activities, and social interaction are the most important ecosystem cultural service elements in these parks.

4.5. The Impact of Landscape Element Types on Residents’ Emotions in Green Spaces

Based on the classification of the five types of green spaces from Table 2, the review data were categorized to calculate sentiment scores for perceptions of nine landscape element types within each park type alongside the overall sentiment scores of the review texts. These data were then imported into Origin 2021 to create a correlation heatmap using the PCCS. The heatmap illustrates the correlations between visitors’ sentiment scores and their perceptions of different landscape element types within each park. It also shows the correlations between pairs of landscape elements. The results are presented in Figure 9.
The results show the following:
Comprehensive Parks: All nine landscape element types are significantly positively correlated with visitor sentiment scores. Water bodies exhibit the highest correlation, reaching moderate strength, followed by supporting facilities. Other elements show smaller but consistent positive correlations. Regarding the pairwise correlations, all are significantly positive, with buildings and rest facilities exhibiting very strong correlations. Buildings and structures are strongly correlated, while moderate correlations are found between rest facilities and roads and squares and between recreational facilities.
Cultural Relics Parks: All nine landscape element types are significantly positively correlated with visitor sentiment scores, with water bodies showing the strongest correlation. Other elements exhibit weaker yet positive correlations. Pairwise correlations reveal that certain relationships—between vegetation and buildings, structures, and rest facilities; animals and buildings, structures, recreational facilities, and rest facilities; and structures and recreational facilities and supporting facilities—are not significant. However, all others show a significant positive correlation, with strong correlations between buildings and rest facilities and moderate correlations between vegetation and animals, vegetation and roads and squares, water bodies and buildings, and buildings and structures.
Ecological Parks: All nine landscape element types show significant positive correlations with visitor sentiment, with water bodies showing strong correlations and supporting facilities having moderate ones. Regarding pairwise correlations, roads and squares, buildings, and structures do not show significant correlations with vegetation or animals. Structures also show no significant correlations with recreational facilities, supporting facilities, or rest facilities. However, buildings and rest facilities have very strong correlations, recreational facilities and rest facilities are strongly correlated, and moderate correlations are found between roads and squares and buildings, roads and squares and structures, and buildings and structures.
Recreational Parks: All nine landscape element types are significantly positively correlated with visitor sentiment scores, with water bodies and supporting facilities showing moderate correlations. In terms of pairwise correlations, no significant correlations were found between buildings and structures, vegetation and animals, recreational facilities and supporting facilities, rest facilities and structures, or roads and squares and structures. However, all other pairs show a significant positive correlation. Specifically, moderate correlations are observed between vegetation and animals, animals and water bodies, roads and squares and buildings, roads and squares and rest facilities, buildings and recreational facilities, and buildings and rest facilities.
Community Parks: All nine landscape element types are significantly positively correlated with visitor sentiment scores, with recreational facilities showing strong correlations and animals and water bodies showing moderate correlations. The pairwise correlations were more complex. Vegetation, animals, and the various facility elements do not show significant correlations. However, very strong positive correlations were observed between water bodies and structures, buildings and rest facilities, and recreational facilities and rest facilities. Additionally, there are strong positive correlations between vegetation and buildings, while moderate positive correlations were noted between vegetation and roads and squares, vegetation and rest facilities, and water bodies and recreational facilities. Interestingly, animals and buildings showed a significant negative correlation.

4.6. The Impact of External Environmental Factors on Resident Emotions

The PCCS between the perceived sentiment scores for the four external environmental factors and the sentiment scores from resident reviews were calculated, as shown in Table 8.
The results show a significant positive correlation between all four external environmental factors and the sentiment scores in resident reviews. Weather changes exhibited a moderate correlation, while the other factors showed weak correlations.

4.7. SEM of the Impact of Green Space Elements on Residents’ Mental Health

The geographic data obtained through satellite remote sensing and the perceived sentiment scores of each element derived from social media data through text and sentiment analysis were imported into Amos. A SEM was constructed to examine the relationships between twenty-three elements and residents’ emotions. After continuous testing and modification of the model, the final result is presented in Figure 10 and Table 9.
The standardized path coefficients indicate the strength and direction of the effects, with p-values less than 0.05 showing significant correlations. Table 9 shows that external characteristics, ecosystem cultural services, landscape elements, and environmental factors all have significant positive effects on residents’ emotions, supporting the validity of the model’s hypotheses.
The fit indices suggest that the model has a good fit (Table 10). The chi-square to degrees-of-freedom ratio (χ2/df) is 2.136, below the threshold of 3. All goodness-of-fit indices (GFI, IFI, TLI, CFI) are greater than 0.90, and the RMSEA index is 0.046, which is below the 0.08 threshold for a good fit.
The results indicate that all factors—external characteristics, ecosystem cultural services, landscape elements, and environmental factors—positively influence residents’ emotions. Among them, landscape elements had the most significant positive impact on residents’ emotions, followed by ecosystem cultural services. External characteristics and environmental factors had the smallest but still positive impact.

5. Discussion

5.1. The Overall Impact of Various UGS Elements on Residents’ Mental Health

Based on the four initial hypotheses and the SEM from Section 4.7, the following conclusions can be drawn:
(1)
The external characteristics of UGS have a positive impact on residents’ emotions.
(2)
The ecosystem cultural services of UGS positively affect residents’ emotions.
(3)
The landscape elements in UGS positively influence residents’ emotions.
(4)
The environmental factors of UGS positively affect residents’ emotions.
(5)
Among these, landscape elements exert the strongest positive influence on residents’ emotions, while the other three factors have relatively small and similar levels of impact.
The analysis in Section 4.5 on the influence of landscape elements reveals that water bodies consistently exhibit a significant strong positive correlation with residents’ emotions, with correlation values of 0.53, 0.62, 0.65, 0.59, and 0.57, respectively. This suggests that water bodies in UGS greatly contribute to the positive psychological well-being of visitors. Additionally, supporting facilities show a moderate, significant positive correlation, with correlation coefficients of 0.40, 0.37, 0.42, 0.43, and 0.39, respectively. This highlights that well-planned and sufficient supporting infrastructure improves the visitor experience and boosts positive emotions. In terms of the relationship between various elements, natural and artificial landscape features generally show weak or insignificant correlations across the five types of parks. This finding suggests the need to balance natural and artificial features in urban park design to enhance the overall experience. A clear zoning strategy is recommended to delineate park areas: natural preservation zones emphasizing indigenous ecosystems, recreational zones, and cultural display zones. Ecological walkways and bridges facilitate ecological connectivity between functionally differentiated areas. Shanghai Houtan Park demonstrates the effectiveness of this strategy through spatially stratified zones, including wetland conservation areas (recording a 23% biodiversity increase), agricultural–industrial heritage sites, and post-industrial revitalization spaces. This integrated approach allows for both ecosystem protection and cultural continuity in urban park design.
The other three aspects have a similar impact on residents’ mental well-being. As shown in Section 4.4, the composition of cultural service elements in a park’s ecosystem generally determines aspects such as park type, nature, and target population. Therefore, when considering the impact of green space parks on the mental health of nearby residents, urban planners need to thoroughly investigate the needs of the surrounding population and design a reasonable ecosystem cultural service structure.
According to Section 4.3, factors such as transportation accessibility, green vegetation availability, and irregular park boundaries significantly influence residents’ emotions, with their p-values all below 0.001. This suggests that urban planners should improve transportation links to ensure broad resident access to parks. Moreover, the park’s greening and layout should promote relaxation, as greater green space availability promotes visitor comfort, while irregular park shapes can inspire a sense of exploration. Planners must address the needs of those with limited mobility by providing more accessible infrastructure to ensure that everyone can enjoy the park’s greenery.
Section 4.6 highlights that weather conditions have the most substantial influence on residents’ emotions, with an r-value of 0.402, compared to the other three factors, all below 0.400. Clear and mild weather tends to encourage outdoor activities, while poor weather can negatively affect visitors’ mood even before a visit begins. This suggests that park management should monitor weather conditions and implement appropriate measures, such as improving drainage systems in areas prone to heavy rainfall and adding shelters or pavilions for rain protection. In regions with intense sunlight, shaded paths can offer relief to visitors. Given the increasing frequency of extreme weather due to climate change, park managers can use smart city monitoring platforms to access real-time and forecast meteorological data, enabling them to prepare effectively for heatwaves, flooding, and other weather-related emergencies. A collaborative study by Western Sydney University and Eratos presents a scalable and sustainable smart city framework capable of reducing urban heat island temperatures by up to 10 °C. This system utilizes IoT (Internet of Things) sensors integrated with machine learning algorithms to dynamically optimize irrigation patterns and water resource allocation in response to real-time meteorological data.

5.2. The Impact of Various Elements in Different Types of Green Spaces on Residents’ Mental Health

Based on the findings from Section 4.1 regarding the distribution of average sentiment among different park types, comprehensive parks are more likely to elicit positive emotions in residents, with approximately 86.37% of reviews in such parks expressing positive sentiments. Combined with the results from Section 4.5, it becomes clear that comprehensive parks are unique in that all landscape element sentiment scores, resident sentiment scores, and pairwise correlations between element sentiment scores are significantly positive. This indicates that comprehensive parks offer the most channels through which landscape elements positively influence visitors’ psychological well-being. Consequently, comprehensive parks should be designed with a holistic approach, considering various elements to maximize positive impacts. The findings from Section 4.4 show that comprehensive parks provide a wide range of ecosystem cultural services, catering to diverse groups and serving a broad population. Among the five park types, comprehensive parks best meet residents’ outdoor exercise needs, far surpassing the other types, with the perceived frequency of outdoor workouts reaching 0.395—significantly exceeding the average value of 0.205. This may be a key reason why comprehensive parks foster positive emotions effectively. Therefore, when designing parks, planners should prioritize convenient exercise facilities, such as jogging paths and fitness areas.
In contrast, cultural relics parks show the most varied sentiment distribution, with a significant difference between the maximum and minimum sentiment values. As highlighted in Section 4.4, cultural relics parks offer a limited range of ecosystem cultural services, focusing on history and culture, while other services fall below average. This likely results in a smaller audience compared to other park types, leading to a wider range of emotional experiences among visitors. To avoid this, parks should avoid over-relying on a single type of ecosystem cultural service. In addition to providing the primary service, parks should incorporate other ecosystem services to attract a wider range of visitors.
Section 4.5 reveals that roads and squares are significantly positively correlated with natural landscape elements in comprehensive parks, cultural relics parks, recreational parks, and community parks, though this correlation is not significant in ecological parks. Similarly, buildings are positively correlated with natural landscape elements in comprehensive and community parks but not in cultural relics, ecological, or recreational parks. This indicates that the design of artificial elements, such as roads and buildings, should vary by park type. In ecological parks, where there are fewer man-made structures, roads and squares should be limited to avoid damaging natural vegetation. In contrast, in parks with more artificial design elements—such as comprehensive, cultural relics, recreational, and community parks—roads and squares can complement the greenery, enhancing the aesthetic. Comprehensive and community parks, often located near urban areas with surrounding buildings, should be designed to harmonize the built environment with natural vegetation. In contrast, cultural relics, ecological, and recreational parks, typically farther from urban centers to offer an escape from city life, should incorporate fewer building elements. These findings suggest that future park management or redevelopment should be tailored to each park’s conditions, with attention to the arrangement of multiple elements to create a unique, cohesive landscape. High Line Park, located in New York City, has repurposed an abandoned elevated railway into a vibrant urban green corridor through the creation of a landscaped pathway with integrated recreational zones. Its design limits hardscape elements while incorporating green infrastructure and curated artistic interventions that preserve the site’s industrial heritage. This adaptive reuse revitalizes the surrounding urban fabric and fosters both community engagement and sustainable development.

5.3. Limitations and Prospects

Since this study sources data from Dianping.com, its findings are limited by the user demographics of the website and may not fully represent all visitors to UGSs. Future research could address this limitation by collecting data on multiple online platforms to obtain comprehensive information about residents [74].
Due to data limitations, this study could not consider all potential factors influencing residents’ mental health, such as age [75], gender, income [76], education level [77], seasonal changes [78], and plant structure [79]. Future studies may address these limitations by combining offline survey data to obtain relevant information about the impact of age, income, education level, seasonal changes, plant structure, etc., on residents’ mental health.
While big data from social media provide unprecedented scale for analyzing user perceptions, their use may carry risks of selection bias. Individuals posting online reviews may systematically differ from general park visitors in demographics or behavior, influenced by access to technology, motivation to engage publicly, or cultural norms. Consequently, emotional responses captured via this approach may inadequately represent all UGS users. Future studies should integrate traditional methods (e.g., surveys) to validate findings, thereby balancing big data advantages with enhanced generalizability.

6. Conclusions

Social media data, characterized by their large volume, real-time nature, dynamic features, and objectivity, provide a timely and accurate reflection of public emotional experiences. They offer an effective, low-cost method for studying UGS. The authors use social media data to measure residents’ emotions and attitudes toward park elements and establish a comprehensive framework to evaluate the impact of UGS elements on residents’ mental health.
Using data from Dianping.com, this study applied text and sentiment analysis techniques, along with geographic information data, to construct a SEM in Amos software. The model examines the relationships between five external characteristic elements, five types of ecosystem cultural services, nine types of landscape elements, four environmental factors, and visitors’ emotions. Findings indicate that all four aspects—external characteristics, ecosystem cultural services, landscape elements, and environmental factors—have a positive influence on residents’ emotions. Among these, landscape elements exert the strongest positive impact, with a coefficient of 0.73, while the other three aspects demonstrate comparable influence levels, among which ecosystem cultural services show a slightly higher coefficient of 0.52, followed by the remaining two aspects with values of 0.32 and 0.31, respectively.
To enhance residents’ recreational experiences and promote their physical and mental health, the following UGS design recommendations are proposed: (1) In park design, designers should give priority to water features, vegetation, and other natural elements and balance them with artificial elements to fulfill the psychological needs of residents to connect with nature. (2) Before park design, planners should assess the needs of residents to ensure that the cultural services of the park’s ecosystem meet their needs. (3) Parks should provide spaces for physical activities, such as walking, running, tai chi, and square dancing. (4) Designers should integrate barrier-free facilities so that people with mobility disabilities, such as parents pushing strollers, wheelchair users, and those with walking aids, can also enjoy the park, improving park inclusivity and green space equity. (5) Designers should consider the characteristics of tourists in the design of roads and squares. It is crucial to design different types of parks to enhance the positive emotions of tourists. (6) Planners should give priority to native plants suitable for local natural conditions to reduce the risk of invasive alien species. In addition, designers should consider biodiversity and provide habitats for birds and other wildlife. (7) Designers should consider multiple sensory experiences in the design process, including sight, smell, hearing, touch, and taste, and other elements. It is feasible to add interactive landscape designs that meet residents’ aesthetic preferences to enhance positive emotions. (8) Designers should ensure that UGSs provide adequate emergency shelter for residents in the event of extreme weather and natural disasters, such as heatwaves, typhoons, heavy rains, or earthquakes [80,81].
This study provides valuable insights for park managers, planners, landscape designers, and policymakers, offering theoretical implications for the science-based planning and management of UGS and parks.

Author Contributions

Writing—original draft preparation, Z.L. and T.D.; writing—review and editing, Z.L. and T.D.; software, Z.L.; formal analysis, Z.L.; data curation, Z.L.; conceptualization, T.D.; methodology, T.D. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Humanities and Social Sciences Youth Foundation, Ministry of Education of the People’s Republic of China. Title: Research on typical urban green space landscape design method based on the relationship between ecosystem cultural services and residents’ health. Grant number: 23YJCZH039.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

All data sources are described in Section 3.1. of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Makutėnienė, D.; Staugaitis, A.J.; Makutėnas, V.; Grīnberga-Zālīte, G. The impact of economic growth and urbanisation on environmental degradation in the Baltic states: An extended Kaya identity. Agriculture 2023, 13, 1844. [Google Scholar] [CrossRef]
  2. Izakovičová, Z.; Petrovič, F.; Pauditšová, E. The impacts of urbanisation on landscape and environment: The case of Slovakia. Sustainability 2021, 14, 60. [Google Scholar] [CrossRef]
  3. Xu, Z.; Gahr, M.; Xiang, Y.; Kingdon, D.; Rüsch, N.; Wang, G. The state of mental health care in China. Asian J. Psychiatr. 2022, 69, 102975. [Google Scholar] [CrossRef] [PubMed]
  4. Zhang, F.; Qian, H. A comprehensive review of the environmental benefits of urban green spaces. Environ. Res. 2024, 252, 118837. [Google Scholar] [CrossRef]
  5. Qi, Y.; Chen, Q.; Lin, F.; Liu, Q.; Zhang, X.; Guo, J.; Qiu, L.; Gao, T. Comparative study on birdsong and its multi-sensory combinational effects on physio-psychological restoration. J. Environ. Psychol. 2022, 83, 101879. [Google Scholar] [CrossRef]
  6. Schebella, M.F.; Weber, D.; Schultz, L.; Weinstein, P. The wellbeing benefits associated with perceived and measured biodiversity in Australian urban green spaces. Sustainability 2019, 11, 802. [Google Scholar] [CrossRef]
  7. Fisher, J.C.; Irvine, K.N.; Bicknell, J.E.; Hayes, W.M.; Fernandes, D.; Mistry, J.; Davies, Z.G. Perceived biodiversity, sound, naturalness and safety enhance the restorative quality and wellbeing benefits of green and blue space in a neotropical city. Sci. Total Environ. 2021, 755, 143095. [Google Scholar] [CrossRef]
  8. Bakhtsiyarava, M.; Ju, Y.; Moran, M.; Rodríguez, D.A.; Dronova, I.; Delclòs-Alió, X.; Moore, K.; Castillo-Riquelme, M.; Anza-Ramirez, C. Associations between urban greenspace and depressive symptoms in Mexico’s cities using different greenspace metrics. Appl. Geogr. 2024, 164, 103219. [Google Scholar] [CrossRef]
  9. Reyes-Riveros, R.; Altamirano, A.; De La Barrera, F.; Rozas-Vásquez, D.; Vieli, L.; Meli, P. Linking public urban green spaces and human well-being: A systematic review. Urban For. Urban Green. 2021, 61, 127105. [Google Scholar] [CrossRef]
  10. Ha, J.; Kim, H.J.; With, K.A. Urban green space alone is not enough: A landscape analysis linking the spatial distribution of urban green space to mental health in the city of Chicago. Landsc. Urban Plan. 2022, 218, 104309. [Google Scholar] [CrossRef]
  11. Bressane, A.; da Cunha Pinto, J.P.; Goulart, A.P.G.; de Castro Medeiros, L.C. Which dimensions of nature contact in Urban Green Spaces most significantly contribute to mental wellbeing? A multidimensional analysis in Brazilian metropolitan cities. Health Place 2024, 89, 103335. [Google Scholar] [CrossRef] [PubMed]
  12. Zhang, R.; Peng, S.; Sun, F.; Deng, L.; Che, Y. Assessing the social equity of urban parks: An improved index integrating multiple quality dimensions and service accessibility. Cities 2022, 129, 103839. [Google Scholar] [CrossRef]
  13. Liu, B.; Zheng, S.; Zhang, L.; Liu, J.; Fu, T.; Hao, R.; Yin, M. Identification and Analysis of Potential Open-Sharing Subjects of Unit-Affiliated Green Spaces in Shanghai Based on POI Data. Land 2023, 12, 2162. [Google Scholar] [CrossRef]
  14. Sergiacomi, C.; Vuletić, D.; Paletto, A.; Barbierato, E.; Fagarazzi, C. Exploring National Park Visitors’ Judgements from Social Media: The Case Study of Plitvice Lakes National Park. Forests 2022, 13, 717. [Google Scholar] [CrossRef]
  15. Li, J.; Gao, J.; Zhang, Z.; Fu, J.; Shao, G.; Zhao, Z.; Yang, P. Insights into citizens’ experiences of cultural ecosystem services in urban green spaces based on social media analytics. Landsc. Urban Plan. 2024, 244, 104999. [Google Scholar] [CrossRef]
  16. Ge, F.; Tang, G.; Zhong, M.; Zhang, Y.; Xiao, J.; Li, J.; Ge, F. Assessment of ecosystem health and its key determinants in the middle reaches of the yangtze river urban agglomeration, china. Int. J. Environ. Res. Public Health 2022, 19, 771. [Google Scholar] [CrossRef]
  17. Delgado-Serrano, M.M.; Melichová, K.; Mac Fadden, I.; Cruz-Piedrahita, C. Perception of green spaces’ role in enhancing mental health and mental well-being in small and medium-sized cities. Land Use Policy 2024, 139, 107087. [Google Scholar] [CrossRef]
  18. Wang, R.; Browning, M.H.; Kee, F.; Hunter, R.F. Exploring mechanistic pathways linking urban green and blue space to mental wellbeing before and after urban regeneration of a greenway: Evidence from the Connswater Community Greenway, Belfast, UK. Landsc. Urban Plan. 2023, 235, 104739. [Google Scholar] [CrossRef]
  19. Reid, C.E.; Rieves, E.S.; Carlson, K. Perceptions of green space usage, abundance, and quality of green space were associated with better mental health during the COVID-19 pandemic among residents of Denver. PLoS ONE 2022, 17, e0263779. [Google Scholar] [CrossRef]
  20. Zhang, J.; Liu, Y.; Zhou, S.; Cheng, Y.; Zhao, B. Do various dimensions of exposure metrics affect biopsychosocial pathways linking green spaces to mental health? A cross-sectional study in Nanjing, China. Landsc. Urban Plan. 2022, 226, 104494. [Google Scholar] [CrossRef]
  21. Yigitcanlar, T.; Kamruzzaman, M.; Teimouri, R.; Degirmenci, K.; Alanjagh, F.A. Association between park visits and mental health in a developing country context: The case of Tabriz, Iran. Landsc. Urban Plan. 2020, 199, 103805. [Google Scholar] [CrossRef]
  22. Yuen, H.K.; Jenkins, G.R. Factors associated with changes in subjective well-being immediately after urban park visit. Int. J. Environ. Health Res. 2020, 30, 134–145. [Google Scholar] [CrossRef] [PubMed]
  23. Singh, K.K.; Rho, S.; Singh, A.; Sergei, C. Big data analytics and knowledge discovery for urban computing and intelligence. Complex Intell. Syst. 2024, 10, 1–2. [Google Scholar] [CrossRef]
  24. Farkas, J.Z.; Hoyk, E.; de Morais, M.B.; Csomós, G. A systematic review of urban green space research over the last 30 years: A bibliometric analysis. Heliyon 2023, 9, e13406. [Google Scholar] [CrossRef] [PubMed]
  25. Ghermandi, A.; Sinclair, M. Passive crowdsourcing of social media in environmental research: A systematic map. Glob. Environ. Chang. 2019, 55, 36–47. [Google Scholar] [CrossRef]
  26. Toivonen, T.; Heikinheimo, V.; Fink, C.; Hausmann, A.; Hiippala, T.; Järv, O.; Tenkanen, H.; Di Minin, E. Social media data for conservation science: A methodological overview. Biol. Conserv. 2019, 233, 298–315. [Google Scholar] [CrossRef]
  27. Walter, M.; Bagozzi, B.E.; Ajibade, I.; Mondal, P. Social media analysis reveals environmental injustices in Philadelphia urban parks. Sci. Rep. 2023, 13, 12571. [Google Scholar] [CrossRef]
  28. Shi, L.; Maruthaveeran, S.; Yusof, M.J.M.; Dai, C. Exploring Factors Influencing Recreational Experiences of Urban River Corridors Based on Social Media Data. Appl. Sci. 2024, 14, 4086. [Google Scholar] [CrossRef]
  29. Li, R.; Li, Y.-Q.; Ruan, W.-Q.; Zhang, S.-N.; Wang, M.-Y. Sentiment mining of online reviews of peer-to-peer accommodations: Customer emotional heterogeneity and its influencing factors. Tour. Manag. 2023, 96, 104704. [Google Scholar] [CrossRef]
  30. Zhou, C.; Zhang, S.; Zhao, M.; Wang, L.; Chen, J.; Liu, B. Investigating the dynamicity of sentiment predictors in urban green spaces: A machine learning-based approach. Urban For. Urban Green. 2023, 89, 128130. [Google Scholar] [CrossRef]
  31. Dai, D.; Chen, Y.; Dai, K. Evaluating cultural ecosystem services in China’s modern historic parks: A sentiment computing approach. Urban For. Urban Green. 2024, 95, 128314. [Google Scholar] [CrossRef]
  32. Pickering, C.; Walden-Schreiner, C.; Barros, A.; Rossi, S.D. Using social media images and text to examine how tourists view and value the highest mountain in Australia. J. Outdoor Recreat. Tour. 2020, 29, 100252. [Google Scholar] [CrossRef]
  33. Li, P.; Liu, J.; Luo, A.; Wang, Y.; Zhu, J.; Xu, S. Deep learning method for Chinese multisource point of interest matching. Comput. Environ. Urban. 2022, 96, 101821. [Google Scholar] [CrossRef]
  34. Hou, Y.; Liu, Y.; Wu, Y.; Wang, L. Assessing Inequality in Urban Green Spaces with Consideration for Physical Activity Promotion: Utilizing Spatial Analysis Techniques Supported by Multisource Data. Land 2024, 13, 626. [Google Scholar] [CrossRef]
  35. Kong, L.; Liu, Z.; Pan, X.; Wang, Y.; Guo, X.; Wu, J. How do different types and landscape attributes of urban parks affect visitors’ positive emotions? Landsc. Urban Plan. 2022, 226, 104482. [Google Scholar] [CrossRef]
  36. Şenik, B.; Uzun, O. A process approach to the open green space system planning. Landsc. Ecol. Eng. 2022, 18, 203–219. [Google Scholar] [CrossRef]
  37. De La Iglesia Martinez, A.; Labib, S. Demystifying normalized difference vegetation index (NDVI) for greenness exposure assessments and policy interventions in urban greening. Environ. Res. 2023, 220, 115155. [Google Scholar] [CrossRef]
  38. Jin, B.; Geng, J.; Ke, S.; Pan, H. Analysis of spatial variation of street landscape greening and influencing factors: An example from Fuzhou city, China. Sci. Rep. 2023, 13, 21767. [Google Scholar] [CrossRef]
  39. Long, X.; Lin, H.; An, X.; Chen, S.; Qi, S.; Zhang, M. Evaluation and analysis of ecosystem service value based on land use/cover change in Dongting Lake wetland. Ecol. Indic. 2022, 136, 108619. [Google Scholar] [CrossRef]
  40. Yang, K.; Luo, Y.; Li, M.; Zhong, S.; Liu, Q.; Li, X. Reconstruction of sentinel-2 image time series using google earth engine. Remote Sens. 2022, 14, 4395. [Google Scholar] [CrossRef]
  41. Sun, P.; Song, Y.; Lu, W. Effect of urban green space in the hilly environment on physical activity and health outcomes: Mediation analysis on multiple greenery measures. Land 2022, 11, 612. [Google Scholar] [CrossRef]
  42. Masoudi, M.; Tan, P.Y. Multi-year comparison of the effects of spatial pattern of urban green spaces on urban land surface temperature. Landsc. Urban Plan. 2019, 184, 44–58. [Google Scholar] [CrossRef]
  43. Masoudi, M.; Richards, D.R.; Tan, P.Y. Assessment of the Influence of Spatial Scale and Type of Land Cover on Urban Landscape Pattern Analysis Using Landscape Metrics. J. Geovis. Spat. Anal. 2024, 8, 8. [Google Scholar] [CrossRef]
  44. Wan, J.; Wu, H.; Collins, R.; Deng, K.; Zhu, W.; Xiao, H.; Tang, X.; Tian, C.; Zhang, C.; Zhang, L. Integrative analysis of health restoration in urban blue-green spaces: A multiscale approach to community park. J. Clean. Prod. 2024, 435, 140178. [Google Scholar] [CrossRef]
  45. Zhao, W.; Liu, D.; Niu, J.; He, J.; Xu, F. Spatial Heterogeneity Analysis of the Multidimensional Characteristics of Urban Green Spaces in China—A Study Based on 285 Prefecture-Level Cities. Land 2024, 13, 1050. [Google Scholar] [CrossRef]
  46. Zhang, M.; Wang, J.; Zhang, F. The Influence of Spatial Heterogeneity of Urban Green Space on Surface Temperature. Forests 2024, 15, 878. [Google Scholar] [CrossRef]
  47. Zhang, Y.; Yan, S.; Liu, J.; Xu, P. Popularity influence mechanism of creative industry parks: A semantic analysis based on social media data. Sustain. Cities Soc. 2023, 90, 104384. [Google Scholar] [CrossRef]
  48. Huai, S.; Liu, S.; Zheng, T.; Van de Voorde, T. Are social media data and survey data consistent in measuring park visitation, park satisfaction, and their influencing factors? A case study in Shanghai. Urban For. Urban Green. 2023, 81, 127869. [Google Scholar] [CrossRef]
  49. Zhao, F.; Liu, H. Modeling customer satisfaction and revisit intention from online restaurant reviews: An attribute-level analysis. Ind. Manag. Data Syst. 2023, 123, 1548–1568. [Google Scholar] [CrossRef]
  50. Chen, J.; Becken, S.; Stantic, B. Assessing destination satisfaction by social media: An innovative approach using Importance-Performance Analysis. Ann. Tour. Res. 2022, 93, 103371. [Google Scholar] [CrossRef]
  51. Sun, X.; Liu, Z.; Huo, X. Six-granularity based Chinese short text classification. Ieee Access 2023, 11, 35841–35852. [Google Scholar] [CrossRef]
  52. Yuan, H.; Shen, X.; Kong, G.; Duan, C. Older Adults and Their Families’ Online Reviews of Urban Nursing Homes in China. Gerontologist 2024, 64, gnad134. [Google Scholar] [CrossRef]
  53. Kiplagat, A.K.; Koech, J.K.; Ng’etich, J.K.; Lagat, M.J.; Khazenzi, J.A.; Odhiambo, K.O. Urban green space characteristics, visitation patterns and influence of visitors’ socio-economic attributes on visitation in Kisumu City and Eldoret Municipality, Kenya. Trees Forest. People 2022, 7, 100175. [Google Scholar] [CrossRef]
  54. Koh, Y.F.; Loc, H.H.; Park, E. Towards a “City in nature”: Evaluating the cultural ecosystem services approach using online public participation GIS to support urban green space management. Sustainability 2022, 14, 1499. [Google Scholar] [CrossRef]
  55. Li, J.; Huang, Z.; Zheng, D.; Zhao, Y.; Huang, P.; Huang, S.; Fang, W.; Fu, W.; Zhu, Z. Effect of landscape elements on public psychology in urban park waterfront green space: A quantitative study by semantic segmentation. Forests 2023, 14, 244. [Google Scholar] [CrossRef]
  56. Cao, S.; Shang, Z.; Li, X.; Luo, H.; Sun, L.; Jiang, M.; Du, J.; Fu, E.; Ma, J.; Li, N. Cloudy or sunny? Effects of different environmental types of urban green spaces on public physiological and psychological health under two weather conditions. Front. Public Health 2023, 11, 1258848. [Google Scholar] [CrossRef]
  57. Ma, S.; Wang, B.; Liu, W.; Zhou, H.; Wang, Y.; Li, S. Assessment of street space quality and subjective well-being mismatch and its impact, using multi-source big data. Cities 2024, 147, 104797. [Google Scholar] [CrossRef]
  58. Huai, S.; Van de Voorde, T. Which environmental features contribute to positive and negative perceptions of urban parks? A cross-cultural comparison using online reviews and Natural Language Processing methods. Landsc. Urban Plan. 2022, 218, 104307. [Google Scholar] [CrossRef]
  59. Li, Y.; Afzaal, M.; Yin, Y. A lexicon-based diachronic comparison of emotions and sentiments in literary translation: A case study of five Chinese versions of David Copperfield. PLoS ONE 2024, 19, e0297101. [Google Scholar] [CrossRef]
  60. Li, S.; Zhu, B.; Zhang, Y.; Liu, F.; Yu, Z. A Two-Stage Nonlinear User Satisfaction Decision Model Based on Online Review Mining: Considering Non-Compensatory and Compensatory Stages. J. Theor. Appl. El. Comm. 2024, 19, 272–296. [Google Scholar] [CrossRef]
  61. Wu, K.; Wu, J.; Li, Y. Mining typhoon victim information based on multi-source data fusion using social media data in China: A case study of the 2019 Super Typhoon Lekima. Geomat. Nat. Hazards Risk 2022, 13, 1087–1105. [Google Scholar] [CrossRef]
  62. Saura, J.R.; Palacios-Marqués, D.; Ribeiro-Soriano, D. Exploring the boundaries of open innovation: Evidence from social media mining. Technovation 2023, 119, 102447. [Google Scholar] [CrossRef]
  63. Xu, J.; Guo, D.; Xu, J.; Luo, C. How Do Multiple Actors Conduct Science Communication About Omicron on Weibo: A Mixed-Method Study. Media Commun. 2023, 11, 306–322. [Google Scholar] [CrossRef]
  64. Wang, C.; Zhu, R.; Zhong, J.; Shi, H.; Liu, C.; Liu, H.; Tan, B.; Xiang, L.; Xiang, R.; Ye, X. Smellscape Characteristics of an Urban Park in Summer: A Case Study in Beijing, China. Sustainability 2023, 16, 163. [Google Scholar] [CrossRef]
  65. Xue, K.; Yu, K.; Zhang, H.; Liang, X. Research on health promotion strategies of public recreation space in the coastal area of Qingdao City Center, China. Sustain. Energy Technol. 2022, 52, 102144. [Google Scholar] [CrossRef]
  66. Xiang, Y.; Hedblom, M.; Wang, S.; Qiu, L.; Gao, T. Indicator selection combining audio and visual perception of urban green spaces. Ecol. Indic. 2022, 137, 108772. [Google Scholar] [CrossRef]
  67. van den Heuvel, E.; Zhan, Z. Myths about linear and monotonic associations: Pearson’sr, Spearman’s ρ, and Kendall’s τ. Am. Stat. 2022, 76, 44–52. [Google Scholar] [CrossRef]
  68. Arif, M.; Petrosillo, I.; Changxiao, L. Effects of changing riparian topography on the decline of ecological indicators along the drawdown zones of long rivers in China. Front. For. Glob. Chang. 2024, 7, 1293330. [Google Scholar] [CrossRef]
  69. Yin, Z.; Zhang, Y.; Zhang, R.; Chen, G.; Cong, Y.; Ma, K. Structure of an urban green space indirectly affects the distribution of airborne particulate matter: A study based on structural equation modelling. Urban For. Urban Green. 2022, 72, 127581. [Google Scholar] [CrossRef]
  70. He, D.; Miao, J.; Lu, Y.; Song, Y.; Chen, L.; Liu, Y. Urban greenery mitigates the negative effect of urban density on older adults’ life satisfaction: Evidence from Shanghai, China. Cities 2022, 124, 103607. [Google Scholar] [CrossRef]
  71. Lefcheck, J.S. piecewiseSEM: Piecewise structural equation modelling in r for ecology, evolution, and systematics. Methods Ecol. Evol. 2016, 7, 573–579. [Google Scholar] [CrossRef]
  72. Zhao, Y.; van den Berg, P.E.; Ossokina, I.V.; Arentze, T.A. How do urban parks, neighborhood open spaces, and private gardens relate to individuals’ subjective well-being: Results of a structural equation model. Sustain. Cities Soc. 2024, 101, 105094. [Google Scholar] [CrossRef]
  73. Wilson, B.; Neale, C.; Roe, J. Urban green space access, social cohesion, and mental health outcomes before and during COVID-19. Cities 2024, 152, 105173. [Google Scholar] [CrossRef]
  74. Owuor, I.; Hochmair, H.H.; Paulus, G. Use of social media data, online reviews and wikipedia page views to measure visitation patterns of outdoor attractions. J. Outdoor Recreat. Tour. 2023, 44, 100681. [Google Scholar] [CrossRef]
  75. Li, H.; Browning, M.H.; Dzhambov, A.M.; Zhang, G.; Cao, Y. Green space for mental health in the COVID-19 era: A pathway analysis in residential green space users. Land 2022, 11, 1128. [Google Scholar] [CrossRef]
  76. Wang, P.; Han, L.; Hao, R.; Mei, R. Understanding the relationship between small urban parks and mental health: A case study in Shanghai, China. Urban For. Urban Green. 2022, 78, 127784. [Google Scholar] [CrossRef]
  77. Mendoza, H.; Rodriguez-Loureiro, L.; Gadeyne, S.; Lefebvre, W.; Vanpoucke, C.; Casas, L. Urban green spaces and suicide mortality in Belgium (2001–2011): A census-based longitudinal study. Environ. Res. 2023, 216, 114517. [Google Scholar] [CrossRef]
  78. Zhou, Y.; Yang, L.; Yu, J.; Guo, S. Do seasons matter? Exploring the dynamic link between blue-green space and mental restoration. Urban For. Urban Green. 2022, 73, 127612. [Google Scholar] [CrossRef]
  79. Cao, S.; Song, C.; Jiang, S.; Luo, H.; Zhang, P.; Huang, Y.; Yu, J.; Li, K.; Li, N.; Guo, B. Effects of Urban Greenway Environmental Types and Landscape Characteristics on Physical and Mental Health Restoration. Forests 2024, 15, 679. [Google Scholar] [CrossRef]
  80. Dong, T.; Feng, C.R.; Yue, B.G.; Zhang, Z.D. An Evaluation Model of Urban Green Space Based on Residents’ Physical Activity. Sustainability 2024, 16, 4220. [Google Scholar] [CrossRef]
  81. Dong, T.; Zhong, Q.K.; Yue, B.G. How Green Space Justice in urban built-up areas affects public mental health: A moderated chain mediation model. Front. Public Health 2024, 12, 1442182. [Google Scholar] [CrossRef] [PubMed]
Sustainability 17 03465 g001
Figure 1. Research area.
Figure 1. Research area.
Sustainability 17 03465 g001
Sustainability 17 03465 g002
Figure 2. Hypothetical model of the mechanism through which park elements influence residents’ mental health.
Figure 2. Hypothetical model of the mechanism through which park elements influence residents’ mental health.
Sustainability 17 03465 g002
Sustainability 17 03465 g003
Figure 3. Research Process.
Figure 3. Research Process.
Sustainability 17 03465 g003
Sustainability 17 03465 g004
Figure 4. Heatmap of average sentiment scores for the studied parks.
Figure 4. Heatmap of average sentiment scores for the studied parks.
Sustainability 17 03465 g004
Sustainability 17 03465 g005
Figure 5. Positive and negative sentiment distribution in park reviews.
Figure 5. Positive and negative sentiment distribution in park reviews.
Sustainability 17 03465 g005
Sustainability 17 03465 g006
Figure 6. Scatter plot of sentiment scores by park type. Note: SS refers to sentiment score.
Figure 6. Scatter plot of sentiment scores by park type. Note: SS refers to sentiment score.
Sustainability 17 03465 g006
Sustainability 17 03465 g007
Figure 7. Distribution of average sentiment scores by park type.
Figure 7. Distribution of average sentiment scores by park type.
Sustainability 17 03465 g007
Sustainability 17 03465 g008
Figure 8. Perception frequency of cultural service elements in different types of park ecosystems. (a) Comprehensive parks. (b) Cultural relics parks. (c) Ecological parks. (d) Recreational parks. (e) Community parks.
Figure 8. Perception frequency of cultural service elements in different types of park ecosystems. (a) Comprehensive parks. (b) Cultural relics parks. (c) Ecological parks. (d) Recreational parks. (e) Community parks.
Sustainability 17 03465 g008
Sustainability 17 03465 g009
Figure 9. Correlation heatmap between the perceived sentiment scores of landscape element types and visitor sentiment scores across different park types. (a) Comprehensive parks. (b) Cultural relics parks. (c) Ecological parks. (d) Recreational parks. (e) Community parks. Note: Veg refers to vegetation; Ani refers to animals; WB refers to water bodies; RS refers to roads and squares; Bui refers to buildings; Str refers to structures; RecF refers to recreational facilities; SF refers to supporting facilities; ResF refers to rest facilities; SS refers to sentiment score.
Figure 9. Correlation heatmap between the perceived sentiment scores of landscape element types and visitor sentiment scores across different park types. (a) Comprehensive parks. (b) Cultural relics parks. (c) Ecological parks. (d) Recreational parks. (e) Community parks. Note: Veg refers to vegetation; Ani refers to animals; WB refers to water bodies; RS refers to roads and squares; Bui refers to buildings; Str refers to structures; RecF refers to recreational facilities; SF refers to supporting facilities; ResF refers to rest facilities; SS refers to sentiment score.
Sustainability 17 03465 g009
Sustainability 17 03465 g010
Figure 10. SEM of the relationship between green space elements and residents’ mental health.
Figure 10. SEM of the relationship between green space elements and residents’ mental health.
Sustainability 17 03465 g010
Table 1. Spatial characteristics of urban parks.
Table 1. Spatial characteristics of urban parks.
Spatial Characteristics of Green SpacesMeasurement IndicatorsBibliography
AvailabilityNDVI[37,41]
ConnectivityArea-weighted mean euclidean nearest neighbor distance (ENN_AM)[42,43]
Size areaArea[44,45]
Boundary shapeArea-weighted mean shape index (SHAPE_AM)[43,46]
Table 2. Categories of UGS.
Table 2. Categories of UGS.
Type of ParkDefinitionQuantities
Comprehensive ParksThese green areas are content-rich, ideal for various outdoor activities, and offer full recreational and support services.14
Cultural Relics ParksThese areas are green parks primarily made up of significant relics and their surrounding environment, emphasizing relic preservation and exhibition. They hold exceptional historical and cultural importance, serving cultural, recreational, and additional purposes.11
Ecological ParksEcological parks offer forest tourism and nature landscape tours to the public, with extensive woodland as their primary feature.21
Recreational ParksRecreational parks feature children’s areas, zoos, botanical gardens, and sports fields, each designed to support play and typically serve a single function.13
Community ParksThese parks are self-contained, featuring basic recreational and service amenities, primarily for nearby community residents to engage in daily recreational activities near other service zones.3
Table 3. External characteristics of urban parks.
Table 3. External characteristics of urban parks.
External CharacteristicsInstructionsTotal Number of Words
Transport conditionAccess to the park’s surroundings, access to public transport248
Table 4. Types of cultural services in urban parks.
Table 4. Types of cultural services in urban parks.
Types of Cultural ServicesInstructionsTotal Number of Words
Recreational activitiesAmong such services provided by the park, residents can engage in various types of cultural and recreational activities, including performances, exhibitions, etc.47
Aesthetic appreciationIn this type of service provided by the park, residents can perceive, understand, and evaluate the landscape, layout, and elements, including marveling at and appreciating the various types of scenery.81
Outdoor workoutsAmong such services provided by the park, residents can engage in various types of physical exercise and fitness activities, including running, playing ball games, etc.64
History and cultureIn such services provided by the park, residents can experience various elements reflecting historical changes, cultural heritage, and regional characteristics, including museums, ancient buildings, etc.74
Social interactionIn this type of service provided by the park, residents can engage in various forms of communication, sharing, and other social behaviors with each other, including chatting, taking photos, etc.34
Table 5. Types of landscape elements in urban parks.
Table 5. Types of landscape elements in urban parks.
Landscape Element TypesInstructionsTotal Number of Words
Natural landscape elementsVegetationVarious types of plants in the park, including trees, shrubs, herbs, aquatic plants, etc. 115
AnimalsAnimals in the park, including birds, mammals, insects, aquatic animals, etc.117
Water bodiesWater bodies in parks, including lakes, ponds, streams, fountains, waterfalls, etc.208
Artificial landscape elementsRoads and squaresIncluding lanes, walkways, trails, staging plazas, recreation plazas, cultural plazas, etc.72
BuildingsIncluding service buildings, public administration facilities, tourist buildings, etc.167
StructuresIncluding sculptures, fountains, and other landscape features and structures, such as steps and walls.82
Recreational facilitiesIncluding amusement parks, aquariums, ballparks, etc.55
Supporting facilitiesIncluding transport facilities, public service facilities, catering, etc.149
Rest facilitiesIncludes seating, gazebos, campsites, etc.97
Table 6. External environmental factors in urban parks.
Table 6. External environmental factors in urban parks.
External Environmental FactorsInstructionsTotal Number of Words
Weather conditionsWeather changesCloudy, sunny, rainy, snowy, etc.39
Extreme weatherRainstorms, snowstorms, typhoons, thunderstorms, etc.77
Circadian conditionsDaytimeTime words for daytime.35
Night timeTime words for evening.16
Table 7. Pearson correlation between the external characteristics of UGS and average resident sentiment scores.
Table 7. Pearson correlation between the external characteristics of UGS and average resident sentiment scores.
External Characteristicsrp
Transport condition0.597 ***0.000
Availability0.578 ***0.000
Connectivity0.283 *0.026
Size area0.301 *0.017
Boundary shape0.698 ***0.000
* p ≤ 0.05 *** p ≤ 0.001.
Table 8. PCCS between external environmental factor sentiment scores and resident review sentiment scores.
Table 8. PCCS between external environmental factor sentiment scores and resident review sentiment scores.
External Environmental Factorsrp
Weather conditionsWeather changes0.402 ***0.000
Extreme weather0.325 ***0.000
Circadian conditionsDaytime0.352 ***0.000
Night time0.350 ***0.000
*** p ≤ 0.001.
Table 9. Summary of regression coefficients for the model.
Table 9. Summary of regression coefficients for the model.
DirectionStandardized CoefficientsUnstandardized CoefficientsS.E.C.R.p
External characteristics
→Residents’ mental health		0.3180.0420.0162.5720.010
Types of cultural services
→Residents’ mental health		0.6530.2190.2192.9890.003
Landscape Element Types
→Residents’ mental health		0.3850.0810.0814.7470.000
External environmental factors
→Residents’ mental health		0.1910.0880.0882.1800.029
The symbol “→” denotes either a regression effect or a measurement relationship.
Table 10. SEM fit indices.
Table 10. SEM fit indices.
Indexχ2/dfRMSEAGFIIFITLICFI
Recommended value<3<0.08>0.9>0.9>0.9>0.9
Model value2.1360.0460.9180.9700.9600.970
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Li, Z.; Dong, T. Exploring the Mental Health Benefits of Urban Green Spaces Through Social Media Big Data: A Case Study of the Changsha–Zhuzhou–Xiangtan Urban Agglomeration. Sustainability 2025, 17, 3465. https://doi.org/10.3390/su17083465

AMA Style

Li Z, Dong T. Exploring the Mental Health Benefits of Urban Green Spaces Through Social Media Big Data: A Case Study of the Changsha–Zhuzhou–Xiangtan Urban Agglomeration. Sustainability. 2025; 17(8):3465. https://doi.org/10.3390/su17083465

Chicago/Turabian Style

Li, Zhijian, and Tian Dong. 2025. "Exploring the Mental Health Benefits of Urban Green Spaces Through Social Media Big Data: A Case Study of the Changsha–Zhuzhou–Xiangtan Urban Agglomeration" Sustainability 17, no. 8: 3465. https://doi.org/10.3390/su17083465

APA Style

Li, Z., & Dong, T. (2025). Exploring the Mental Health Benefits of Urban Green Spaces Through Social Media Big Data: A Case Study of the Changsha–Zhuzhou–Xiangtan Urban Agglomeration. Sustainability, 17(8), 3465. https://doi.org/10.3390/su17083465

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop