Evaluating Green Campus Environments in Chinese Universities from Subjective Perceptions: A Textual Semantic and Importance–Performance Analysis Through a Satisfaction Survey

Abstract

University campuses play a crucial role in sustainable development; however, the current green campus evaluation systems tend to focus primarily on the physical environment and building technology, often overlooking user perception and the impact of these factors on the execution of green campus strategies. Starting with an examination of the connotation and evolution of green campuses, we derived relevant indicators of campus subjective perceptions from both domestic and international green campus evaluation systems. We collected user feedback through satisfaction questionnaires and text data on the green campuses of nine representative Chinese universities. Factor analysis was used to establish the correlations between campus planning and subjective perceptions across six key areas. This research applied importance–performance analysis (IPA) to assess the prioritization of each green campus indicator, integrating it with textual semantic analysis to better understand the perceptions and attitudes of campus users toward green campus development. The findings suggest that the objectives of a green campus cannot be fully achieved using only technical or physical evaluation criteria. Instead, combining subjective feedback with quantitative indicators forms the foundation for effective strategy development. This study also found that users were more concerned about the design of details related to learning, living, entertainment, and recreation than the broader green campus planning decisions made by planners and decision-makers. Focusing on user perception and balancing scientific planning with public participation can help achieve the ultimate goal of green campus planning and design, adhering to a human-centered approach.

1. Introduction

In the context of global sustainability, university campuses, often referred to as miniature cities, encompass complex systems involving land, buildings, activities, transportation, and energy use. They play a vital role in sustainable development [1]. As future decision-makers, university students’ awareness of green lifestyles will be crucial to building an environmentally conscious society; hence, the construction of sustainable campuses for universities has garnered significant attention [2]. In 1972, the United Nations Conference on the Human Environment first introduced the concept of the “green campus” [3]; later, the Talloires Declaration broadened the concept of sustainability to encompass all aspects of teaching and research [4]. Subsequently, the Association for the Advancement of Sustainability in Higher Education (AASHE) and the Universitas Indonesia launched the STARS and GreenMetric green campus evaluation systems, which set out evaluation indicators in terms of the “hardware” or physical space and the “software” or educational operations [5,6]. In the People’s Republic of China, according to the 2019 national standard of green campus assessment, a green campus is defined as “a harmonious environment that provides teachers and students with safe, healthy, suitable, and efficient learning and living spaces, optimizes resource conservation, protects the environment, reduces pollution, and imparts educational value to students”. This definition broadens the concept of a green campus by incorporating aspects such as resource conservation, environmental education, and campus space creation. This definition explains the green campus concept concerning resource conservation, environmental education, and space creation, emphasizing the “greening” of the built environment in the physical space dimension, while also indicating the importance of raising the awareness of teachers and students through the “greening” of education and research in the behavioral consciousness dimension.

Most of the current research on green campuses has largely focused on the physical environment, such as energy-efficient buildings, ecological landscape design, and sustainable transport systems [7,8,9,10,11]. Green campus evaluations have predominantly relied on objective indicators related to physical infrastructure, including energy consumption rates, green space designs, and environmental impacts [12,13,14]. This focus has often led to an emphasis on the physical environment. Research comparing the green campus policies of various Canadian universities using the STARS evaluation system found that these institutions typically prioritized physical environment indicators over socio-cultural factors and user perception [15]. While enhancing the objective evaluation system has contributed to a better research environment for green campuses, a persistent issue in green design is the misalignment between evaluation indicators and actual energy consumption. The concept of a green campus should not be confined to the energy performance of campus buildings or limited to landscaping or land development [16,17,18]. Research on LEED buildings on university campuses revealed that there is no direct correlation between LEED ratings and actual energy savings. In fact, several platinum-rated buildings consume more energy than average buildings and fail to achieve the expected performance in terms of user satisfaction [19,20]. This phenomenon highlights that objective technical indicators alone cannot fully assess whether a campus is truly “green and sustainable”, necessitating the inclusion of human factors in green design considerations.

Recent research has highlighted an important insight in the field of green design: human behavior is the ultimate determinant of the sustainable performance of physical spatial environments. It is essential for green design to align with the laws of human behavior and consider how to guide people toward making the right decisions [21]. A randomized questionnaire conducted among university students in Hawaii and Alabama, USA, revealed that students’ sense of personal responsibility and involvement in sustainable campus environments significantly influenced their level of awareness and participation in green campus initiatives [22]. Furthermore, research on student perceptions and the intensity of use of campus green spaces showed that those who frequently utilized a specific green area were more attuned to the overall green space, demonstrating the substantial impact of behavioral habits on sustainability [23]. However, the majority of studies at this stage are conducted from the perspectives of designers and managers, exploring the key factors that influence the construction of green campuses through the integration of an objective indicator evaluation system [24,25,26]. While these studies provide valuable recommendations for policymakers to manage the development of green campuses through a top-down approach, they overlook the perspectives and insights of the actual campus users themselves. The perceptions of students and teachers, as direct service recipients of the campus, directly influenced their satisfaction with and engagement in the campus environment [27]. Measuring users’ attitudes toward green campuses can reveal areas for improvement in campus planning and construction, providing essential insights for evaluating the effectiveness of green campus strategies [28,29]. The current research lacks the evaluation of green campus planning and construction from the user’s perspective. How to transform the subjective perceptions of teachers and students regarding the green campus space into quantifiable and analyzable data indicators has become an urgent issue for informing green campus construction.

Questionnaires remain a valid method for quantifying subjective perceptions in fields such as psychology, social sciences, and urban research [30,31,32]. Integrating importance–performance analysis (IPA) with questionnaire-scale data offers a systematic framework for the user-centered evaluation of green campus planning. IPA is a tool used to assess user satisfaction by comparing users’ perceived importance of various environmental aspects with their actual performance [33]. IPA has been widely employed for satisfaction evaluations in tourism, residential areas, greenways, ecological service systems, and other fields [34,35,36,37]. It helps planners assess the significance of campus environmental indicators and users’ satisfaction levels. Although evaluation systems like STARS and GreenMetric provide benchmarks for objective assessment, they may not fully capture the student experience on campus. Recommendations from teachers and students better reflect users’ perceptions, needs, and expectations in terms of a green campus, compensating for the limitations of objective assessments. Textual semantic analysis, which thoroughly examines sentiment and semantic information from suggestion texts to reveal hidden patterns, emotions, and themes, has gained widespread use in fields like urban parks, public participation, and neighborhood spatial evaluation [38,39,40]. Text semantic analysis can provide a more profound understanding of the experiences and expectations of students and teachers regarding the campus environment, offering an innovative approach to evaluating green campus satisfaction.

The rapid construction of extensive “university cities” in China is slowing down, and sustainable campus development is shifting from energy-saving to green campuses. Maintaining and renewing long-established campuses while solving issues such as unconventional expansion, excessive land use, landscape commercialization, and opulent buildings has gradually become the focus of sustainable development [41]. In recent years, most sustainable development efforts in relation to Chinese universities have primarily focused on ecological aspects, often neglecting humanistic considerations, community interactions, and economic factors [14]. In comparison to the green campuses in Europe and the United States, which focus more on the publicity and education of sustainable concepts, Chinese university campuses still emphasize the demonstration of eco-technology and energy management of facilities, with a limited understanding of the interrelations among campus elements and the enhancement of how different campus elements interact and how user–environment interactions can enhance campus vitality [3].

In light of the global transformation of the concept of green campus construction in universities, scientifically evaluating the effectiveness of green campus initiatives has become a critical topic. This research aims to answer the following questions by assessing the subjective perceptions of students and teachers regarding the green campus environments across nine universities in China: (a) How satisfied are students and teachers with the current green campuses? (b) What are the key factors influencing the satisfaction with the green campuses? (c) Which aspects of green campus planning need improvement? The findings of this study are expected to offer design and management recommendations for university administrators and urban planners, contributing to the enhancement of green campus strategies and furthering theoretical research in the field.

2. Method

2.1. Selection of Sample Universities

While there has been substantial practice-based research on university sustainability programs, there is still a lack of categorical comparisons and subjective perception feedback on the policies and programs implemented by multiple universities. Conducting cross-sectional comparative research among universities can help explore the conceptualization and execution of their green initiatives, enabling each institution to address its shortcomings through mutual improvement [42]. The selection of university samples plays a crucial role in shaping the research method. The subjective perceptions and evaluative judgments of teachers and students regarding the campus environment are influenced by factors such as the university type, university level, campus size, construction period, urban location, and building climate demarcation. This study considers these factors to ensure that the selected green campuses of Chinese universities are representative and diversified.

First of all, this study considers the background of the universities. Comprehensive universities tend to focus on campus space development through the integration of multiple disciplines, agricultural and forestry universities prioritize environmental configurations suited to the specific needs of their fields, while science and engineering universities design campuses around experimental facilities and research sites. The campus design brought about by the university type significantly affects the sensitivity and clarity of the campus environment perceived by students and teachers. Therefore, this study selects comprehensive universities, agricultural and forestry universities, and science and engineering universities to reflect the different characteristics of the built environment. In addition, this study takes the university level into account. The sample includes six Double First-Class universities and three general universities, covering institutions of various levels and rankings. Second, this study considers the campus size and construction period. Universities with areas ranging from 33.3 to 301.3 ha and population densities ranging from 82.58 to 538.79 persons/ha are selected, creating a spatial scale gradient from small intensive to large comprehensive. The sample includes universities established over varying time frames, ensuring a balanced representation of campuses featuring both mature landscapes and contemporary plans. Third, given the diverse geography of China and the influence of climate change on environmental perception, the sample is based on the building climate and physical environment design basis for thermal zoning (Standard of climatic regionalization for architecture GB 50178-93) [43], covering the typical building climate zones, including the severe cold regions in the northwest and northeast, the cold regions in the west and north, and the hot summer and cold winter regions. Based on these factors, nine universities are selected for the study, with detailed information provided in Figure 1.

2.2. Questionnaire Design and Implementation

Enhancing users’ participation and incorporating their voices into campus planning can be effectively achieved by assessing students’ perceptions and evaluations of the university’s green initiatives [44]. The questionnaire survey remains an essential method for obtaining subjective perception data. The questionnaire consisted of three sections: the first section gathered respondents’ basic information; the second section focused on core items, using indicators from the green campus evaluation system to assess satisfaction; and the third section included an open-ended question-and-answer segment aimed at collecting respondents’ evaluations and suggestions regarding green campus planning and development. The green campus satisfaction evaluation was based on indicators including green buildings, landscape environment, convenient transportation, and community connectivity as outlined in the assessment standard for green campus (GB/T51356-2019) [45] (

Table 1. Relevant items in the criteria corresponding to subjective perception.

Category	Clause	Scoring Item	Specification	Evaluation Basis	Evaluation Method	Value
Green building thermal environment	3.7	Indoor hot and humid environments	Thermal comfort of teaching rooms	Evaluation standard for indoor thermal environment in civil buildings GB/T50785 [46]	Simulation analysis report; comfort satisfaction practical research	12
			Thermal comfort of administrative offices
			Thermal comfort of dormitories
	3.11	Campus heat island environments	Outdoor shade area	Clauses	Relevant design documents; Site verification and measurement	6
			Coefficient of reflection of solar radiation from road and roof surfaces
Green building light environment	1.3 *	Lighting for public buildings and dormitories	Meet the requirements of relevant daylighting standards	Code for design of dormitory building JGJ36 [47]	Relevant design documents; daylighting simulation; site verification and measurements	—
	3.6	Lighting for main function rooms	Compliance area for teaching rooms	Standard for daylighting design of buildings GB50033 [48]	Relevant design documents; site verification and measurement	11
			Compliance area for administrative offices
			Compliance area for dormitories
Green building sound environment	3.1 *	School environmental noise	Noise compliance	Environmental quality standard for noise GB3096 [49]	Site verification and measurement	—
	3.2 *	Building noise and sound insulation	Noise level and sound insulation performance are up to standard	Comply with the Code for design of sound insulation of civil buildings GB50118 [50]	Site verification and measurement	—
	3.5	Indoor noise and sound insulation	Noise level of main function rooms	Better than the lowlimit standard of the Code for design of sound insulation of civil buildings GB50118 [50]	Relevant design documents; site verification and measurement	12
			Building components and airborne sound insulation performance
			Impact sound insulation performance of building floor slabs
Green building wind environment	1.3 *	Indoor ventilation	Public buildings and dormitories	Ensure ventilation requirements	Relevant design documents; site verification and measurement	—
	1.8	Campus wind environment	Typical winter wind speed and pressure	Clauses	Wind environment simulation	8
			Transitional season, summer wind speed and pressure
Landscape environment	3.10	Surface water quality	Environmental quality of surface water	Environmental quality standards for surface water GB3838 [51]	Relevant monitoring materials	9
	3.12	Campus greening	Selection of native plants	Clauses	Relevant design documents; site verification	9
			Tree configuration
			Vertical, rooftop and other greening methods
Convenient transportation	1.11 1.12	Public transport connections Rational parking planning	Distance from entrances to stations	Clauses	Relevant design documents; site verification	9 10
			Number of public transport routes
			Convenient pedestrian access
			Bicycle parking facilities
			Manner, place, efficiency of use, design and temporary parking of motor vehicles
			Open and shared motor vehicle parking
Community connectivity	1.13	Public service resource sharing	Centralized design of public service functions	Clauses	Relevant design documents; site verification	10

Notes: * Indicates a control item.

). This evaluation also referenced the American STARS system and the Indonesian GreenMetric system, both of which are closely related to the student’s learning and living experiences. The satisfaction for each indicator was rated on a five-point scale: very satisfied, satisfied, average, dissatisfied, and very dissatisfied. The analysis was performed using the Likert scale cumulative method, which assigns values of 5, 4, 3, 2, and 1 to each of the five levels of response (Appendix A).

A total of 2533 questionnaires were collected for this study. After eliminating invalid questionnaires with elevated vacancy rates, excessive answer repetition, and contradictions, 1985 valid and trustworthy surveys were retained. To assess the accuracy, stability, and validity of the questionnaires, this study subjected the retained valid questionnaires to reliability and validity tests. The results showed that the Cronbach’s alpha was 0.866 > 0.7, and the KMO value was 0.910 > 0.7, indicating a high level of reliability and validity in terms of the data. To avoid common method biases in the questionnaire, this study conducted a further Harman’s single-factor test, revealing that the explained variance for each component was below 40%, suggesting the absence of significant common method biases.

This study performed an exploratory factor analysis on the pre-set indicators of the questionnaire to assess the reasonableness of its design. Maximum likelihood rotation and eigenvalues greater than 1 were used to extract potential factors. Nineteen indicators were finally selected for evaluating green campus satisfaction in this study, after removing indicators that contributed to low Cronbach’s alpha values, low factor loadings, and cross-loadings. These indicators include the proximity to functional zones, area ratio of functional zones, spatial clarity, spatial recognition, parking bay availability, parking area, walking safety perception, greening proportion, greening hierarchy, waterscape quality, general landscape, classroom lighting, classroom ventilation, classroom noise, hot and humid environments, participating in environmental protection lectures, involvement in energy-saving and environmental activities, university hospital facilities, and campus awareness.

2.3. Data Analysis

2.3.1. Exploratory Factor Analysis and Weight Calculation

An exploratory factor analysis of the 19 indicators assessing satisfaction with green campus planning facilitated the examination of the covariance relationships among the satisfaction indicators included in the questionnaire, thereby categorizing them into several potentially influential factors and evaluating the construct validity [52]. Maximum likelihood rotation and eigenvalues greater than 1 were used to extract potential factors. The internal consistency was tested for each influential factor by calculating the Cronbach’s alpha.

Principal component analysis (PCA) is used to determine the factor weights. The principle is to convert a collection of potentially correlated variables (features) into a reduced set of uncorrelated variables (called principal components). Upon dimensionality reduction, the PCA loadings (weights) associated with each principle component can be taken as the value of the importance of the original variables [53]. The weights of the original variables are given by the eigenvectors. If we consider a dataset with variables X₁, X₂, …, X_n, and calculate the first principal component, the weight for each variable X_i in PC1 will be given by the components of the eigenvector corresponding to the largest eigenvalue. Mathematically, if V₁ is the eigenvector corresponding to the largest eigenvalue λ₁, the weight of X_i in the first principal component is as in Equation (1), where V_i1 is the i-th component of the first eigenvector V₁.

Weight of X_i = V_i1

(1)

2.3.2. Importance–Performance Analysis

Importance–performance analysis (IPA) is a sociological research method proposed by Martilla and James [54]. It offers clarity on which service attributes warrant prioritization and is thus extensively utilized in service evaluation of customer satisfaction [54,55]. IPA analyzes the difference between respondents’ perceived importance and the objective performance of the subject, evaluates satisfaction, and formulates optimization and enhancement strategies based on the findings, centering on the creation of the IPA chart depicting the mean values of importance and satisfaction (Figure 2). The indicators in Quadrant I, labeled “Keep up the good work”, exhibit high importance and performance and should be sustained; the indicators in Quadrant II, labeled “Possible overkill”, display high importance and low performance and can be monitored but do not necessitate significant attention; the indicators in Quadrant III, labeled “Low priority”, are defined by low importance and performance, and are often not prioritized; and the indicators in Quadrant IV, labeled “Concentrate here”, demonstrate high importance and low performance, necessitating concentrated attention and enhancement.

Although IPA has been used in a wide range of research areas, determining the optimal placement of the crosshairs in the importance–performance matrix has been a major issue in terms of its application. Traditionally, research has used “scale-centered” IPA, where the crosshairs are simply placed in the middle of the Likert scale used to measure importance and performance [54]. However, the tendency of respondents to give high performance and importance ratings results in the “ceiling effect”, which inflates both the importance and performance scores [56]. “Scale-centered” IPA tends to place indicators in the “Keep up the good work” quadrant, resulting in analytical inaccuracies [57]. To minimize such inaccuracies, this research used a “data-centered” IPA, which places the crosshairs at the mean responses of the importance and performance items measured [58]. This IPA method effectively solves the problem of “ceiling effects” by ensuring that salient attributes are graphed according to their relative importance and performance, which ensures more dispersion of attributes across the four IPA grids and contributes to the accuracy of the analysis [59]. In this research, the weights calculated from the principal component analysis were used as the indicator importance, and the ratings obtained from the questionnaire survey were used as the indicator satisfaction to draw the IPA chart for green campus planning.

2.3.3. Textual Semantic Analysis

Open-ended questions in questionnaires allow respondents to express thoughts and attitudes that structured measures may not capture, facilitating the acquisition of further, more valid information [60]. Research substantiates the significant utility of open-ended questions in surveys [61]. Textual analysis techniques offer a rapid and accurate approach to extracting valid information from open-ended survey questions and have been widely used in fields such as political science, travel behavior, and climate research [62,63,64]. In this study, textual semantic analysis was used to extract suggestions and classifications for a green campus from suggestion text data provided by teachers and students in the questionnaire, including data preprocessing, lexical processing, sentiment analysis, and topic modeling.

Firstly, the data were preprocessed to remove duplicate and irrelevant texts, such as “none”, “no”, “no suggestion”, etc.; subsequently, responses with more than 10 words were filtered out and the text was cleaned (lowercasing, tokenization, removing stopwords, and non-alphabetic words); and finally, 981 effective suggestion texts were collected, with an average length of 29.6 words per text.

Secondly, adverbs, prepositions, conjunctions, auxiliaries, and exclamations, which have no practical meaning, were added to the filtered word list. Then, the preprocessed effective text was imported into the Rost Content Mining 6 software to execute the word segmentation process and the word frequency statistics.

Thirdly, this research used natural language processing (NLP) techniques to analyze the sentiment of the text [65]. Sentiment analysis is a subtask of NLP text classification in which various sentiments are assigned different labels, and then the model learns to classify the text. This research used the BERT model for positive and negative classification of emotions. The BERT model is an unsupervised deep learning model developed by Google AI in 2018. Unlike previous pre-trained models that can only acquire unidirectional contextual information, the BERT model uses the masked language model and a deep bi-directional codec for pre-training, integrating unit word features into the downstream task to generate deep bi-directional linguistic representations that can fuse left and right contextual information [66]. The BERT model is widely used for text prediction due to its superior performance in sentiment analysis. This research used the pre-trained BERT model from the Hugging Face Transformers library to classify the sentiment of teacher and student suggestion texts into positive and negative categories.

Finally, the BERTopic model was introduced for topic modeling of teacher and student suggestion texts. Topic modeling is an unsupervised machine learning technique for determining abstract topics from unstructured textual data, aiming to reveal the underlying topic structure in a collection of texts. BERTopic is a topic model proposed by Maarten in 2022 that uses pre-trained models to create clusters of topics through techniques such as Transformer and c-TF-IDF, preserving key terms in the topic descriptions [67]. Compared to traditional methods using Bag of Words (BoW) techniques like latent Dirichlet allocation (LDA) and probabilistic latent semantic analysis (PLSA), it addresses the limitations of ignoring the word order and the excessive number of words and dimensions that impair the extraction accuracy. Due to the separation of its clustering generation from the creation of topic representations, it supports the automatic identification of the number of topics and is competitive in research involving classical topic models and other clustering methods that follow topic modeling. Thus, it is widely used in fields such as library intelligence, behavioral sciences, and ecological environment [68,69,70]. This research investigated topic clustering using the BERTopic model and a “pre-training + fine-tuning” migration learning method [66]. The operation process is shown in Figure 3.

3. Results

3.1. Green Campus Satisfaction

3.1.1. Characteristics of the Respondents

The data were aggregated by university, with each indicator reflecting the specific conditions of each individual university.

Table 2. Demographic characteristics of respondents from each university.

		BIT	BFU	NCU	NEU	SNNU	XJU	JXNU	KSU	NIT
Gender (%)	Male	40.07	36.30	60.74	69.26	33.44	34.48	52.10	35.25	68.60
	Female	59.93	63.70	39.26	30.74	66.56	65.52	47.90	64.75	31.40
Composition (%)	Undergraduate students	51.92	96.96	93.94	80.53	52.06	98.85	96.13	80.57	88.84
	Master’s degree students	23.00	3.04	5.05	10.31	38.73	1.15	2.26	0.71	10.33
	PhD students	3.14	0.00	0.34	0.38	0.00	0.00	1.61	0.00	0.83
	Teachers	21.95	0.00	0.67	8.78	9.21	0.00	0.00	18.73	0.00
Residence (%)	On campus	88.21	97.79	98.66	93.63	97.14	99.62	95.48	93.64	91.32
	Not on campus	11.79	2.21	1.34	6.37	2.86	0.38	4.52	6.36	8.68

presents the demographic characteristics of the respondents from each university. The respondents consisted of undergraduate students, graduate students, and teachers of varying genders, academic years, and majors, with a distribution of 47.38% males and 52.62% females, which was fairly balanced. Undergraduate students made up the largest group (79.80%), followed by master’s degree students (12%) and teachers (7.42%), while PhD students accounted for the least (0.79%). First- and second-year undergraduate students were over 50% of the total. In terms of residence, the overwhelming majority of respondents (95.14%) reside on campus.

3.1.2. Factor Analysis

To measure the consistency of the questionnaire results, this research used the Cronbach’s alpha as the criterion to test the reliability of the selected indicators and obtained a coefficient of 0.855 > 0.7, indicating a high level of reliability. To determine whether the respondents’ ratings of satisfaction with the green campus were suitable for exploratory factor analysis, this research conducted the KMO test and Bartlett’s test of sphericity. The results showed that the KMO value was 0.89 > 0. 7 and the p-value of Bartlett’s test of sphericity was 0 < 0.05. These findings indicated a high degree of data reliability and suitability for exploratory factor analysis. The Kaiser normalized maximum variance method was used to rotate the factors, and the factor categories were extracted using principal component analysis. The number of potential factors retained was determined according to the gravel plot, resulting in the extraction of six factors from the 19 satisfaction indicators, which explained 61.54% of the total variance (

Table 3. Results of the exploratory factor analysis.

Code	Indicator	Rotated (Varimax) Factor Loading
		F1-Land Use	F2-Planning Pattern	F3-Landscape Environment	F4-Road Transportation	F5-Green Building	F6-Management and Education
LU1	Proximity of functional zoning	0.772	−0.065	0.106	0.161	0.129	0.024
LU2	Area ratio of each functional partition	0.763	0.02	0.144	0.144	0.167	0.113
PP1	Space clarity	0.142	0.78	0.129	0.117	0.028	0.082
PP2	Space recognition	0.239	0.617	0.151	0.115	0.078	0.133
RT1	Parking bay availability (motor vehicles)	0.392	−0.022	0.099	0.526	0.18	0.138
RT2	Parking area (motor vehicles)	0.019	0.026	0.105	0.781	0.157	0.02
RT3	Walking safety perception	0.167	−0.025	0.021	0.665	0.091	0.162
LE1	Greening proportion	0.106	−0.032	0.878	0.054	0.136	0.068
LE2	Greening hierarchy	0.114	−0.008	0.86	0.036	0.194	0.089
LE3	Waterscape quality	0.071	0.071	0.488	0.383	0.242	0.159
LE4	General landscape	0.453	0.017	0.529	0.102	0.227	0.138
GB1	Classroom lighting	0.083	−0.03	0.241	0.026	0.732	0.078
GB2	Classroom ventilation	0.178	0.016	0.14	0.12	0.77	0.111
GB3	Classroom noise	0.084	−0.025	0.094	0.187	0.769	0.053
GB4	Hot and humid environment	0.165	−0.007	0.116	0.176	0.736	0.119
ME1	Participating in lectures on environmental protection	0.445	−0.044	0.138	0.017	0.207	0.59
ME2	Participation in energy-saving and environmental protection activities	0.038	0.069	0.091	0.021	0.07	0.77
ME3	University hospital	0.254	0.203	0.064	0.331	0.201	0.407
ME4	Campus awareness	−0.014	−0.149	0.073	0.229	0.056	0.659
	Standardized Cronbach’s alpha	0.667	0.609	0.778	0.632	0.804	0.617
	Variance explained (%)	10.576	6.655	11.882	9.514	13.874	9.04

Notes: 1. Extraction method: principal component analysis. 2. Rotation method: Kaiser normalized maximum variance method (the rotation has converged after 6 iterations).

).

The first factor (F1) was labeled as “land use”, emphasizing the spatial layout of different functional zones and the proportion of land allocation. It was primarily explained by two indicators: proximity of functional zoning (LU1) and area ratio of each functional partition (LU2). Appropriate land use patterns enhance sustainable and effective land utilization while improving campus accessibility and usability. The second factor (F2) was labeled as “planning pattern”, representing the structural clarity and perceivability of the campus space planning. It included two indicators: spatial clarity (PP1) and spatial recognition (PP2). A clear and easily identifiable campus is more oriented and possesses a heightened sense of place, which contributes to enhancing the user’s visiting experience. The third factor (F3) was labeled as “landscape environment”, focusing on the natural elements and landscape aesthetics of the campus. It comprised four indicators: greening proportion (LE1), greening hierarchy (LE2), waterscape quality (LE3), and general landscape (LE4). Landscaping that merges ecological principles with aesthetic design regulates the microclimate and improves the overall spatial quality of the campus. The fourth factor (F4) was labeled as “road transportation”, highlighting mobility and pedestrian safety. It contained three indicators: parking bay availability (RT1), parking area (RT2), and walking safety perception (RT3). A green campus should consider pedestrian safety and vehicle regulation while encouraging walking and cycling to mitigate carbon emissions and improve campus organization and security. The fifth factor (F5) was labeled as “green building”, concentrating on evaluating the indoor environmental quality of campus buildings. It consisted of four indicators: classroom lighting (GB1), classroom ventilation (GB2), classroom noise (GB3), and hot and humid environment (GB4). Sustainable building design focusing on lighting, ventilation, noise control, humidity, and hot adjustment is essential to reduce energy consumption and improve comfort in educational and residential environments. The sixth factor (F6) was labeled as “management and education”, stressing campus administration and environmental education. It covered four indicators: participating in lectures on environmental protection (ME1), participation in energy-saving and environmental protection activities (ME2), university hospital (ME3), and campus awareness (ME4). A green campus encompasses not only the sustainability of the physical space but also the internalization of environmental awareness. Increased engagement in sustainable development activities and enhanced management will foster the formation and evolution of a green campus culture.

3.1.3. Satisfaction Statistics

The questionnaire provided insights into the satisfaction scores of teachers and students from nine universities regarding their subjective perceptions of green campus construction (Figure 4). The analysis of the satisfaction with green campuses was conducted across six dimensions: land use, planning pattern, landscape environment, road traffic, green building, and management and education, based on the results of the exploratory factor analysis.

The land use section examined the appropriateness of the distance and area ratio of the functional zoning distance and area ratio on each campus. The mean scores across the nine universities were relatively high (3.74, 3.58), suggesting that teachers and students generally perceived the campus land use as reasonable. This finding was highly consistent with previous studies on land resource integration in Chinese universities [71,72], further validating the shared strengths of Chinese universities in intensive land use planning. However, BIT received the lowest scores (3.49, 3.31) in this survey. As a newly constructed suburban campus with extensive land and a well-planned layout, BIT’s lower subjective experience may be attributed to its expansive scale, leading to fragmented functional areas and long travel distances, which negatively affect daily life.

The planning pattern section assessed perceptions of the spatial clarity and recognition of the campus. The “spatial clarity” item measured how easily users navigated the campus, with lower scores indicating higher spatial complexity and a greater likelihood of getting lost. A study of open spaces on Chinese university campuses demonstrated that the design of campuses located in southern cities is relatively focused on responding to natural landforms, exhibiting a rich sense of spatial hierarchy that increases the spatial complexity of users’ subjective perceptions [73]. This may explain the lower scores of NCU and NIT (2.98, 3.18). Similar to the result of Liaoning University’s spatial perception study [74], BFU, SXUN, and KSU received the highest scores in this indicator. All of these campuses are located in northern China and are small in size, suggesting that smaller-scale campuses are easier to navigate, Additionally, it also reflected that the straightforward axial layouts of campuses in northern plains cities effectively enhance the spatial clarity. The “space recognition” item examined whether different campus areas were easily distinguishable by their unique characteristics. In comparison to the research conducted by Yuelu Mountain National University, “spatial recognizability” achieved a high mean score of 3.98 in this survey, suggesting that the majority of campus environments in China possess distinct characteristics and recognizability [75]. NEU, despite an average clarity score, achieved a recognition score of 4.27, whereas NCU and NIT, which excelled in recognition, had lower clarity scores. This finding suggests that high recognition does not necessarily correlate with good spatial clarity. Designers should improve spatial clarity while maintaining spatial interests, ensuring that different areas are both functional and easily navigable to enhance the user experience.

The campus has long served as a green lung within the concrete cityscape. Overall, teachers and students rated the campus landscape environment highly. This section assessed four aspects: greening proportion, greening hierarchy, waterscape quality, and general landscape. Previous studies have shown that Chinese university campuses generally have a high level of landscape satisfaction, especially in terms of the greening [76,77]. In this survey, BFU received the highest score across all four items, averaging 4.0, reflecting an outstanding perception of the campus landscape. This aligns with the university’s academic focus, as BFU boasts the best greenery among the capital universities, providing an exceptional environmental experience for studying and living. The performance of “waterscape quality” differs among the universities. In contrast to the Nanjing Communications Institute of Technology, which exhibited exceptionally high satisfaction [72], BIT attained a score of merely 2.96, indicating a need for improved water management.

In the feedback on the campus green space, a clear distinction emerged between the evaluations of “greening proportion” and “greening hierarchy”. While teachers and students generally perceived the campus as having a high greening proportion, they did not rate the greening hierarchy as equally high, reflecting the current greening design prioritizes “quantity” above “quality”. On the level of subjective perception, the greening proportion was not the most critical factor. Beyond its visual appeal, the effectiveness of campus landscape design lies in its ability to engage users and establish a greening hierarchy that fosters interaction. Additionally, the function of campus green space can be expanded to serve as a platform for ecological sustainability practices and environmental education.

From the score statistics, it was evident that road transportation was one of the most pressing issues across all the survey sections. This category encompassed three items: parking bay, parking area, and walking safety perception. The low mean scores for parking bay (3.37) and parking area (3.13) indicated that motor vehicle parking had a negative impact on the overall campus spatial experience. The mean score for walking safety perception was only 3.4, identical to that of the Wuchang Campus of the Wuhan Institute of Technology, indicating a low level of safety [78]. This indicated that road safety has become a common issue on Chinese university campuses and requires adequate attention. This issue was particularly pronounced in older urban campuses due to the outdated campus planning [71]. For example, the old downtown campus of BFU, where irregular parking, limited parking spaces, and low road safety resulted in consistently low scores across all three indicators (3.32, 2.71, 3.07). Campus parking was often considered a secondary issue in early campus planning in China. As the number of cars soars, issues such as insufficient and poorly designed parking spaces have a negative impact on the campus environments in China [79]. Parking has now become a crucial component of green campus development, and if not addressed, it could have long-term adverse effects on faculty and student spatial perception.

The green building section primarily assessed the subjective perception of various physical attributes of teaching buildings.

Table 4. Environmental evaluation of teaching buildings on the campuses of nine universities.

Name	Improved Campus Environment	Satisfactory Campus Environment
BIT	Air quality	Noise environment
BFU	Noise environment, Air quality	Lighting/Illumination
NCU	Summer thermal environment	Lighting/Illumination, Air quality
NEU	Air quality, Noise environment, Summer thermal environment	Lighting/Illumination
SNNU	Summer thermal environment	Lighting/Illumination
XJU	Noise environment, Summer thermal environment	Air quality
JXNU	Summer/Winter thermal environment	Air quality, Noise environment, Lighting/Illumination
KSU	Summer thermal environment, Air quality	Lighting/Illumination
NIT	Noise environment, Winter thermal environment	Air quality

indicates that the values for lighting, ventilation, noise environment, and thermal environment were evenly distributed, with minimal fluctuation. Teachers and students were generally more satisfied with the teaching buildings. As a key focus in the development of green campuses at domestic universities, the construction of green buildings has made breakthroughs in the design of sound, light, heat, and other building performance and energy management, resulting in better subjective perception [80].

In terms of management and education, participation in environmental protection lectures and energy-saving activities was relatively high. However, the concept of a “Green Campus” remained vague for many respondents. Most indicated a lack of clarity regarding its specific connotation, with an average self-assessment score of 3.06. Previous studies have shown that teachers and students typically recognize the importance of sustainability, although they possess limited comprehension of the notion of a green campus [81]. A survey conducted at Zhongkai University of Agriculture and Engineering and Guangzhou College of South China University of Technology showed that over fifty percent of the students believed they had insufficient information regarding knowledge about a “green campus” [82]. Additionally, there was a general shortage of related courses and activities, highlighting the need for enhanced promotion of the “Green Campus” initiative and strengthened environmental awareness education for both students and teachers [41].

3.2. IPA for Green Campus Planning

Figure 5 shows the IPA scatterplot of all the respondents’ ratings of the green campuses of the nine universities.

Quadrant IV, “Concentrate here”, was the most critical area to address, containing three green campus evaluation indicators: RT1, LE4, and ME3. Most of these indicators pertained to road transportation, landscape environment, and management education. Regarding road transportation, parking bay availability (RT1) was essential for accommodating private vehicles and mitigating campus congestion. However, the low satisfaction scores suggested that unsuitable motor vehicle parking locations, restricted availability, and ineffective utilization of parking spaces had a significant negative impact on the campus environment. The general landscape (LE4) was the most easily perceived and highly valued indicator of a green campus by teachers and students. Enhancing campus aesthetics and visual appeal through effective landscape design plays a key role in green campus planning. In contrast to its importance, the satisfaction evaluation score of the general landscape was inadequate, indicating that the current landscape design had shortcomings and had not fulfilled the expectations of teachers and students. The university hospital (ME3) significantly influenced the teaching and daily life of faculty and students, yet their low satisfaction ratings suggested they required greater attention. Overall, although these indicators were vital for enhancing overall satisfaction with the green campus environment, the actual satisfaction levels were considerably lower than expected. Addressing these shortcomings should be a top priority in campus development.

Quadrant I, “Keep up the good work”, included six green campus evaluation indicators: LU1, LU2, LE1, LE2, LE3, and ME1. Most of these indicators were associated with the land use and landscape environment, indicating that teachers and students generally placed high importance on and expressed strong satisfaction with the functional zoning and greenery and landscape elements. Campus land use planning has garnered significant attention because of the implications for the accessibility and distribution of functional zones, which directly influence the experiences of teachers and students. The elevated satisfaction level further signified that the existing campus planning and design exhibited a considerable degree of reason. In terms of the landscape environment, teachers and students had high requirements for the aesthetics and functionality of green space, water features, and other landscape elements, and the high level of satisfaction highlighted the successful integration of greening and campus planning. Additionally, participation in environmental education activities was highly valued by teachers and students, and the high satisfaction levels suggested strong engagement in and awareness of environmental protection initiatives. Overall, the indicators in Quadrant I reflected high satisfaction levels, reinforcing their essential role in maintaining a positive campus environment. However, the high importance assigned to these indicators also indicated that any decline in their quality would significantly impact overall satisfaction with the green campus, necessitating sustained attention during campus development.

Four evaluation indicators were located in Quadrant II, “Possible overkill”, including PP2, GB1, GB2, and GB3, primarily focused on planning pattern and green buildings. The high satisfaction and low importance of these indicators suggested that teachers and students, as actual users of the green campus, were content but not sensitive to factors related to the macro-design and building technology. Instead, they passively adapted to the designed campus spaces and indoor environments rather than actively engaging with them. This phenomenon was particularly evident in terms of the space recognition (PP2), which recorded the highest satisfaction level but the lowest importance. This imbalance indicated that teachers and students, as high-frequency users of campus space, could easily distinguish different campus zones. Simultaneously, the three indicators (GB1, GB2, and GB3) regarding green buildings indicated that current Chinese green campuses disproportionately emphasized teaching facilities and energy-efficient technologies. While these indicators demonstrated commendable performance, they were not the primary concerns of actual users, despite their previous prominence in the past planning and design of green campuses. Overall, the indicators in Quadrant II were considered secondary areas for maintenance rather than requiring immediate or significant resource allocation for renovation and updates, especially in comparison to indicators in other quadrants.

Six indicators were located in Quadrant III, “Low priority”, including PP1, RT2, RT3, GB4, ME2, and ME4, primarily related to road transportation and management education, both of which showed low levels of importance and satisfaction. The spatial clarity (PP1), which ensures that teachers and students can easily navigate campus zoning and functional facilities, was not perceived as a key factor, although some areas had confusing layouts. Similarly, while the parking area (RT2) and walking safety perception (RT3) scored lower in satisfaction, their relatively low importance suggested that these aspects were not major concerns for enhancing green campus spaces. The hot and humid environment (GB4), although showing low importance similar to other green building indicators, demonstrated poor satisfaction and negatively impacted both thermal comfort and campus sustainability. Compared to participation in environmental protection lectures (ME1), teachers and students showed less interest in energy-saving and environmental protection activities (ME2). Additionally, campus awareness (ME4) performed poorly across the nine universities and failed to attract significant attention. This indicated that raising awareness of sustainability and fostering active engagement in energy-saving initiatives still required substantial effort. Overall, while these indicators could benefit from improvement, addressing them would not significantly enhance the overall satisfaction with the green campus compared to the indicators in other quadrants. Thus, resource allocation should prioritize more critical aspects before focusing on these lower-priority areas.

3.3. Textual Semantics Analysis Based on Suggestions

3.3.1. Sentiment Analysis

The sentiment analysis of the teacher and student suggestion texts from nine universities using the BERT model revealed that positive and negative sentiment statements accounted for 38.12% and 61.88%, respectively, indicating a high proportion of negative sentiment. Further analysis by university (Figure 6) shows that only KSU had a higher proportion of positive statements than negative ones, while NEU had nearly equal proportions. All the other universities exhibited a predominance of negative sentiment. Notably, four universities (BFU, BIT, SXNU, NCU) had a lower percentage of positive statements than the overall average, with NCU recording the highest percentage of negative statements at 82.9%. The high proportion of negative sentiment suggests that most teachers and students who provided suggestions were dissatisfied with the current stage of the green campus planning. Additionally, this negativity may reflect cognitive dissonance, where teachers and students recognize the importance of green campus development but feel that the actual outcomes fall short of their expectations. This mismatch between expectations and reality may manifest as frustration and dissatisfaction in their statements.

Analyzing the five universities with a higher-than-average proportion of positive emotions in the text, it was found that these universities exhibited higher satisfaction with the masterplan, proximity of functional zones, and area ratio of each functional partition. This suggests that well-executed campus planning not only meets the diverse functional needs of teachers and students but also enhances the overall satisfaction and fosters positive emotions toward the green campus. Additionally, teachers and students at these universities expressed greater satisfaction with environmental protection lectures and energy-saving activities, indicating that the organization of such initiatives, along with effective school management, education, and green campus awareness, played a crucial role in promoting emotional positivity.

3.3.2. Keyword Analysis

In order to clarify the key contents of the teachers’ and students’ suggestions for green campus planning, this study used Rost Content Mining software to extract text keywords. The suggestion texts from nine universities underwent word segmentation and frequency analysis, generating a word cloud map to visualize the most frequently mentioned terms, as shown in Figure 7.

The keyword frequencies indicated the focus of students and teachers on various factors. The results showed that among the first 80 high-frequency words, there were 49 nouns, 14 verbs, 14 adjectives, and three adverbs. Nouns accounted for the highest proportion, reflecting four main categories in green campus planning: functional space, infrastructure, transportation, and greening, with terms such as cafeteria, street lighting, roads, and greening. Verbs primarily expressed teachers’ and students’ attitudes toward the construction or renovation of the green campus, including terms like increase and enhance. Adjectives conveyed expectations regarding the convenience, diversity, and satisfaction of the campus environment, with words such as inconvenient, not enough, and unreasonable. Adverbs reflected the urgency and extent of campus renewal, with expressions like as soon as possible and excessively, highlighting the desire of teachers and students for prompt improvements to the campus environment.

To further explore the key factors that teachers and students urgently want to improve, K-means clustering of the keywords based on the word frequency yielded six clusters at level 4 (Figure 8). The high-frequency phrase (Cluster 2) reflected the core concerns of teachers and students and contained the keywords dormitory and greening, indicating that accommodation and the greening environment should be the focus of green campus planning. Three mid-frequency phrases (Cluster 1, 3, and 5) reflected campus planning patterns, land use, and infrastructure issues, respectively, which should be emphasized. Cluster 1 contained the words schools, campus, and environment, which were related to the overall environmental planning of the campus. Cluster 3 emphasized the functional layout of campus planning and included the words supermarket, area, study room, and planning. Cluster 5 mainly reflected problems in public service facilities such as teaching, heating, and lighting, containing words such as classrooms, air conditioner, and street lighting. The low-to-medium-frequency phrase (Cluster 4) reflected the design of public spaces for campus life, study, and leisure, including keywords such as leisure, teaching building, design, plaza, self-study, landscape, hospitals, open space, and bathhouse as well as many other keywords. The low-frequency phrases (Cluster 0) reflected issues related to outdoor and sports venues in campus planning, containing keywords such as outdoor, swimming pool, gymnasium, basketball court, playground, and greenfield. This cluster contained the most vocabulary. The large number of words indicated that, at this stage, general problems existed in these areas across all the universities. However, these issues were not universal, as teachers and students had diverse concerns, and opinions on rectifying the relevant problems varied from one school to another.

3.3.3. Topic Analysis

To further clarify the semantic topics of the teacher and student suggestion texts, combining word meanings and utterance relations based on the keyword frequency, extracting and categorizing topic words with similar meanings or semantic relations can effectively portray the text semantics or features. This study applied BERTopic modeling to the preprocessed text of the teacher and student suggestions and identified 25 topics. Since an excessive number of topics can lead to overly fine granularity, potentially overloading the information, this study utilized the fine-tuning tool in BERTopic to merge similar topics, enhancing the interpretability. Ultimately, nine core topics were refined from the suggestions of the teachers and students. Figure 9 presents the scores of the nine core topic terms, where a higher score indicates a term’s greater representativeness of the topic. The top four representative terms for each topic are automatically used to generate the topic naming.

In contrast to the green campus satisfaction types identified through the factor analysis, BERTopic classified the suggestions from teachers and students into nine principal categories, highlighting that the actual users of the campus have distinct perspectives compared to practitioners in assessing green campus planning. These nine thematic categories reflected the key concerns of the teachers and students, with the highest proportion of suggestions focusing on the campus living environment, facilities, and services. This included dormitories, athletic facilities, logistics services, and healthcare, with a particular emphasis on improving the dormitory environment. Additionally, increasing the green space, enhancing transportation, improving the air quality, accelerating campus construction, and beautifying the environment emerged as major aspirations of teachers and students. These themes collectively reflected their shared vision for sustainable development and a greener campus environment (

Table 5. Text topic classification and content.

Number	Topic Name	Topic Words	Description of Topic Content
Topic 0	Dormitory environment	dormitory, study, campus, rooms, environment	Concerns and needs of teachers and students in campus environment and physical planning. Demonstrates that the dormitory environment is currently the most important concern.
Topic 1	Roads and transportation	parking, street, lights, roads	Concerns of teachers and students regarding transportation and mobility issues. Demonstrates the need for parking facilities and road safety.
Topic 2	Logistics services	cafeteria, supermarket, food, end	Needs of teachers and students in logistics services such as catering and shopping.
Topic 3	Greening Enhancement	green, greenery, greening, increase, little	Desire of teachers and students for more green space on campus and dissatisfaction with the current state of campus greening.
Topic 4	Sports and leisure	swimming, pool, sports, recreational, chairs	Needs of teachers and students for campus recreational activities, sports, and their supporting facilities.
Topic 5	Campus construction	good, pretty, site, finish, soon	The expectations of teachers and students for a speedy improvement of the campus environment and for an aesthetically pleasing and functional plan and design.
Topic 6	Air environment	air, conditioning, units, purification, summer	Concerns of teachers and students regarding indoor environmental comfort and air quality.
Topic 7	Medical facilities	hospital, school, medical, doctors, improved	Needs of teachers and students for enhanced health care services.
Topic 8	Landscape design	trees, plant, lack, landscape, grass	The expectations of teachers and students for improved campus landscaping, especially in terms of plant landscaping.

).

The intertopic distance map further illustrates the interrelationships among the nine topics, where each circle represents a topic, and its size indicates the frequency of the topic across all the documents. Similar topics are positioned closer together in the coordinate system. As shown in Figure 10, the nine topics are grouped into three related clusters. In practice, this grouping allows for addressing issues reflected by highly related topics simultaneously, enabling more effective planning and renovation outcomes.

Group 1 had the largest number of suggested texts and formed the largest cluster. This group was centered on Topic 0, which focused on the accommodation environment and essential life services. A comfortable dormitory environment (Topic 0) was influenced by air-conditioning equipment and indoor air quality (Topic 6), while medical facilities (Topic 5) were closely linked to air quality (Topic 6), the dormitory environment (Topic 0), and logistic services (Topic 2). Poor air quality, suboptimal dormitory conditions, and food safety concerns could negatively impact the health of teachers and students. To address this group, universities should provide diverse and convenient basic services, such as integrating accommodation, healthcare, and catering retail resources, adopting environmentally friendly dormitory designs, and implementing sustainable ventilation systems to maintain indoor air quality, etc. Group 2 was associated with infrastructure development and transportation accessibility. The accessibility of roads and parking areas (Topic 1) and the completion of campus construction (Topic 5) influenced the utilization of sports and leisure facilities (Topic 4). Additionally, an aesthetically pleasing and functional campus design (Topic 5) enhanced the appeal of sports and recreation zones (Topic 4), contributing to the overall green campus planning. To address this group, schools should consider transportation development alongside sports and recreational zoning, focusing on the aesthetic and functional needs of teachers and students while ensuring timely renovations. Group 3 revolved around landscape design and campus greening. Greening enhancement (Topic 3) and landscape design (Topic 8) exhibited a strong correlation, as insufficient green space directly impacted teachers’ and students’ perceptions of the overall campus landscape. To address this group, universities should integrate plant design with the broader landscape layout to create a visually appealing and sustainable campus environment.

Generally speaking, the three groups reflected the concerns and needs of teachers and students regarding the sustainable development of the campus. Group 1 and Group 2 were more closely linked, as well-developed infrastructure and accessible transportation could enhance essential services such as accommodation, catering, and healthcare. Group 3 complemented green campus planning by improving the landscape environment, contributing to a more sustainable and aesthetically pleasing campus.

4. Discussion and Conclusions

This study aimed to evaluate the subjective perceptions of teachers and students at Chinese universities regarding the green campus environment through a satisfaction survey, IPA, and textual semantic analysis. The goal was to elucidate their attitudes and perceptions toward green campus planning and identify areas for improvement in green campus development.

The results show that teachers and students generally perceive campus land use as reasonable, except in new suburban universities, where the excessive spatial scales may lead to issues such as fragmented functional areas and long distances between facilities. Significant differences in planning patterns exist between northern and southern universities, and higher spatial recognizability does not necessarily equate to better spatial clarity. Overall, the campus landscape environments received high scores; however, notable differences emerged in the evaluation of “greening proportion” and “greening hierarchy”. This suggests that a high percentage of green space does not necessarily enhance the perception of greenness. Instead, its effectiveness is closely linked to the uniformity of green space distribution and user accessibility, aspects that are not well captured in current evaluation standards. Road transportation emerged as the most critical issue, with teachers and students expressing concerns about campus parking, road safety, and public transportation connections, which significantly impacted the overall satisfaction. This study also found that green buildings were largely satisfactory in terms of environmental factors such as lighting, ventilation, and thermal comfort. However, management and education were identified as key areas for improvement. There is widespread unawareness of green campus initiatives among educators and students, along with a lack of related programs and activities within universities. Furthermore, the presence or absence of an energy-efficient mindset significantly influences perceptions of campus landscape environments, public spaces, and green buildings.

The IPA prioritizes improvements in green campus planning and construction by identifying indicators that can be enhanced to maximize teacher and student satisfaction. Indicators related to road transportation, landscape environment, green buildings, and management education—such as parking bay availability and general landscape—are located in Quadrant IV. These factors are of high importance for improving the overall satisfaction with the green campus environment but currently receive low satisfaction, making them the highest priority for campus planning improvements. Most indicators related to the landscape environment and land use are positioned in Quadrant I, indicating high satisfaction and high importance. These factors dominate the evaluation and should be maintained as key priorities. Lower-priority indicators related to planning patterns, green buildings, and management education, found in Quadrants II and III, can be improved by reallocating resources to bridge the gap with high-priority areas, ensuring a balanced and effective green campus plan. By understanding the prioritized configuration of indicator improvements and incorporating suggestions from teachers and students, universities can create a more sustainable, functional, and people-centered campus environment.

The semantic analysis of the text reveals that teachers and students are generally dissatisfied with the current state of green campus planning, particularly in terms of the dormitory environment and landscape greening, which are identified as the most urgent areas for improvement. Notably, their suggestions emphasize micro-level factors such as the landscape design, facilities, and transportation accessibility, reflecting a different evaluative perspective from that of planning practitioners. This highlights the importance of user-centered planning and the need for greater public participation in green campus design and decision-making processes. It is also important to acknowledge that most of the proposal texts analyzed in this study represent the concentrated demands of teachers and students, which tend to be more negative in sentiment. This may have introduced bias into the sentiment analysis results. Future research could incorporate in-depth interviews to further enhance the precision and accuracy of the analysis, as well as track the impact of improvements in green campus environment indicators on teacher and student satisfaction to provide a more comprehensive evaluation over time.

In general, this study emphasizes the key role of subjective perception in green campus planning. It reveals that relying solely on data indicators or physical evaluation standards is insufficient to fully achieve the goals of a green campus. Instead, the quantitative performance of green campus planning must be refined into practical strategies to better guide planning and design. The integration of subjective evaluation feedback with quantitative data is essential for developing effective strategies, highlighting the importance of assessing user perceptions of the campus environment. Additionally, in the pre-design phase of the questionnaire, planners tended to evaluate the green campus from a macro, comprehensive perspective, and specialized technical perspective, whereas the proposal texts indicate that teachers and students, as actual users, are more concerned with detailed design elements related to their daily experiences, such as learning, living, recreation, and leisure. The low importance of green buildings and planning pattern indicators in Quadrant II of the IPA diagram further confirms the discrepancy between planners and users in evaluating green campuses. User perception is a decisive factor in assessing the planning and construction of green campuses. Therefore, it is crucial to incorporate the perspectives of teachers and students and to fully leverage public participatory design in future green campus renovations. Finally, while teachers and students may not prioritize macro-level and technical factors such as green buildings, planning patterns, and management and education, universities that score highly in terms of campus green buildings, management, and education tend to exhibit higher overall satisfaction. This suggests that both subjective user perceptions and objective professional assessments are important for measuring green campus indicators, balancing scientific planning with public participation in campus development.

Green campus development is not a short-term integrated construction project but rather a long-term goal focused on sustainable campus development. Given the complex and diverse factors influencing campus spaces, greater attention must be paid to the user perception, energy-saving education, and the establishment of a feedback mechanism alongside objective evaluation systems. This human-centered approach is the ultimate goal of green campus planning and design. Based on the above conclusions, we propose the following improvement strategies for green campus planning in China. (a) Integrate green and sustainable design: expand green spaces while ensuring both functionality and aesthetics. Maintain the prominence of key indicators in Quadrant I, such as greening proportion, greening hierarchy, and waterscape quality, to sustain teacher and student satisfaction. (b) Optimize infrastructure and transportation: address mobility challenges by improving road networks and motor vehicle parking connections. (c) Improve aesthetics and comfort: enhance air quality, seasonal comfort, and spatial aesthetics to create a more pleasant campus environment. (d) Promote diversification of campus life: develop modern sports, recreational, dining, shopping, and medical facilities to enrich campus experiences. (e) Enhance convenience and livability: tackle issues such as noise, litter, and overcrowding while improving dormitory environments to create a more comfortable and functional living space.