Exploring Public Space Satisfaction in Old Residential Areas Based on Impact-Asymmetry Analysis

Abstract

The renovation of public space environments in old residential areas has become the focal point in Chinese urban renewal and promotion of urban sustainable development; thus, an urgent need for research expansion is attached to identifying the environmental factors of public space and their impacts from the perspective of residents’ needs and satisfaction. Taking Hefei, China, as an example, and based on a satisfaction survey of the public space environment in old residential areas, this study discerned critical factors affecting public space satisfaction through gradient-boosting decision trees and impact-asymmetry analysis methods, after which the asymmetric relationship between public space environment factors and resident satisfaction was tested and the priority and goal of public space environment renovation were obtained. The results indicate that the following: (1) Compared with physical environment factors, current social environment factors, including uncivilized behavior, space occupation, and hygiene and cleanliness, exert a greater impact on the overall satisfaction. (2) The hypothesis that there exists a general nonlinear relationship between environmental factors and satisfaction is verified, with all social environment factors proving to be nonlinear and physical environment factors being highly related to social environment creation, such as nursing space for the old and young, reading and learning space, and display and communication space. (3) A priority hierarchy for the improvement of environmental factors should be established, which basically conforms to Maslow’s hierarchy of needs theory. The role of social environment renovation should be highlighted, along with the necessity to bolster community governance and public space management. At the level of the physical environment, more space should be available for the activities of residents, display and communication, and reading and learning. This study can provide a corresponding theoretical basis and planning inspiration for meeting the multiple needs of residents for public space, improving life happiness, and promoting the sustainable development of cities in the renovation of public space environments in old residential areas.

1. Introduction

Urban renewal and reconstruction remain the core issues in the research and practice of global urban planning [1]. To promote comprehensive green transformation and sustainable development in cities, China has entered a critical period of urban renewal. As a state plan, the renovation of old residential areas has become the focal point of Chinese urban renewal. The majority of physical entities that are most closely related to residents’ daily life in residential areas are “public spaces”, which, as a medium of public communication, play an important role in enhancing community cohesion [2], strengthening community identity [3], promoting the integration of social relations [4], and inheriting collective memory [5]. However, there currently exist many problems in the public spaces of old residential areas, including insufficient space, single functions, a lack of supporting facilities, inadequate maintenance, etc. Therefore, public space forms the core of the renovation of old residential areas. In order to promote the efficient renovation of public space in old residential areas, the government and planners must figure out the key spatial attributes that affect resident satisfaction. How to determine what environmental factors are the salient ones affecting resident satisfaction and how to scientifically distinguish the primary and secondary relations of influencing factors have become the key scientific problems to be tackled urgently in the renovation of old residential areas.

Satisfaction is the criterium to measure the consistency between expectation and reality [6]. The concept of satisfaction came from marketing management, which was developed and matured in the fields of economics and sociology. After the 1950s, research on satisfaction extended to the field of urban planning, forming the concept of resident satisfaction. Resident satisfaction is the satisfaction generated by residents in a specific place [7], and it is determined by the gap between the actual living environment and the expected living environment [8]. Public space is an important communication medium in cities, and the optimization of the public space environment can improve the well-being of residents [9]. Studying resident satisfaction with public space can promote the improvement and sustainable development of urban public space environments. Public spaces in cities, such as squares [10], parks [11], and residential areas [12], are our main research objects. The factors that affect satisfaction from public space [13,14,15] and the construction and quantification of an evaluation system [16,17] are the primary research contents. The Chinese government is currently investing heavily in renewing old residential areas to improve resident happiness [18]. In order to renovate the public space in old residential areas and enhance resident happiness, the relevant departments or enterprises of urban planning must understand the main environmental factors that affect residents’ satisfaction as well as dissatisfaction. As government funds and planning resources are always limited, an analysis of influencing factors regarding satisfaction becomes the mainstream of research, and planners are interested in determining the critical factors that affect resident satisfaction and the primary–secondary relationship among factors in need of improvement. When exploring specific influencing factors, Hur et al. emphasized the impact of building density and vegetation coverage rate on residential satisfaction, considering the influence of both factors on perception and evaluation [19]. Using the factor analysis and logit model, K. Lovejoy et al. found that residents of both suburban and traditional neighborhoods consider safety and an attractive appearance to be very important, and for residents of traditional neighborhoods, liveliness and diversity really matter [20]. Lee et al. explored the effects of objective community features and the subjective feelings of residents on resident satisfaction [21]. Sun et al. used structural equation models to examine the correlation between life satisfaction and perceived environmental factors in residential areas as well as the workplace [22]. Chen et al. employed the fuzzy evaluation method to study resident satisfaction in rural concentrated residential areas and found that the cost of living and income are the most important factors affecting resident satisfaction [23]. Hong studied Caoyang New Village in Shanghai and found the positive impact of per capita living space made on satisfaction regarding residential space as well as the environment [24]. Lv et al. used the factor analysis and importance–performance analysis (IPA) methods to jointly evaluate satisfaction and importance in resident satisfaction, and they concluded that the renovation of old residential areas is a priority [25]. Deng et al. used a multi-level linear model to analyze the influence of travel times, neighborhood relations, and degree of concern on resident satisfaction, and they put forward treatment suggestions from the perspective of residents’ demands [26]. Zhang et al. applied the structural equation model (SEM) and found that renovation projects such as external building structures, internal functional space, and public infrastructure had a positive effect on residents’ life satisfaction [27].

Additionally, most of the early studies on resident satisfaction were based on the premise that there exists a linear correlation between environmental factors and resident satisfaction [28]. However, in recent studies, Cao [18,29,30] and Dong et al. [31] studied neighborhood satisfaction and pedestrian satisfaction by using the methods of IAA and GBDT, which all indicate that the previous assumption is invalid. If the relationship between environmental factors and satisfaction is non-linear, then an assumption in favor of the linear relationship may produce biased estimates that can affect planners’ judgment of the true relationship [18]. Based on data collected from Suzhou, China, Fan et al. used GBDT to illustrate the nonlinear relationship between neighborhood factors and life satisfaction [30]. Cao et al. collected data from six old residential areas in Harbin, China, determined the factor structure of neighborhood satisfaction by factor analysis and IAA, and found most neighborhood factors exhibit a non-linear relationship with neighborhood satisfaction, such as housing and public facilities, neighborhood facilities, and the social environment [18]. Dong used the method of IAA to compare the pedestrian satisfaction of residents living in open and gated communities in Harbin, China, and found that most of the walkability factors have a nonlinear influence on pedestrian satisfaction, among which social interaction is the focus of improvement in gated communities [31]. Cao et al. promoted the development of satisfaction research with the linear hypothesis relaxed, and more research is needed to evaluate the universality of nonlinear relationships [18].

At present, at the residential area level, most of the nonlinear relationships affecting satisfaction are studied from the scale of neighborhood and community, and the important influencing factors that are found are often broad, such as housing and public facilities, neighborhood facilities, and the social environment [18]. When it comes to specific measures for the renovation of old residential areas, more detailed funding arrangements and renovation goals are needed, causing the necessity for more detailed studies. With an investigation of 10 old residential areas in Hefei, China, this study tackles three research questions by applying the methods of impact-asymmetry analysis (IAA) and gradient-boosting decision trees (GBDT): (1) What are the leading factors that affect resident satisfaction of the public space environment in old residential areas? (2) Is there a nonlinear relationship between environmental factors and resident satisfaction? (3) In order to boost resident satisfaction, what elements need to be improved first? And to what extent? This study contributes to examining the asymmetric impact of public space factors on resident satisfaction, rectifies the stereotype that all factors share the same degree of influence in previous research and planning practice, provides a further directional reference regarding how to determine the sequential planning of public space environment factors in the practice of old residential renewal for maximizing resident satisfaction, and achieves the optimal allocation of limited resources in a time sequence in the grand public affairs of old residential renewal.

2. Materials and Methods

2.1. Study Area

Hefei, as the capital city of Anhui Province, China, has enjoyed the fastest Gross Domestic Product (GDP) growth rate in China for the past decade, developing from a “small county” with an economic aggregate of less than CNY 100 million in the early years of the founding of New China to the rising sub-center of the Yangtze River Delta world-class urban agglomeration. As is shown in Figure 1, the urban expansion of Hefei also reflects the rapid progress of China’s urbanization, and its population density, residential space structure, and expansion patterns differ from those of mega-cities, such as Beijing and Shanghai. Nevertheless, its residential construction and renovation of old residential areas are representative to a large extent; thus, focusing on the public space of old residential areas in Hefei can effectively supplement previously existing research.

The public space of old residential areas covered by this study refers to those built in Hefei from 1980 to 2000, mainly consisting of collective self-built housing (unit-based welfare housing, collective housing, and staff dormitories), housing reform (public housing purchased by urban workers), early commercial housing, and relocation support housing. Public space refers to all outdoor public activity areas that can be freely accessed and used, excluding the indoor space of residential buildings. It incorporates public areas in buildings, traffic roads, the greening landscape, infrastructure, and hard activity sites (such as plazas, playgrounds, fitness spaces, etc.), which are main sites of daily leisure activities for residents. Therefore, the satisfaction level of residents also responds to their specific needs for the renovation of public space. To ensure that the selected samples can basically reflect the construction features of old residential areas in Hefei and the existing problems regarding public space use, we first constructed a geographic database of residential areas in Hefei (Figure 1), based on which we further screened out the residential areas built from 1980 to 2000. Considering the heterogeneous impact of construction time, geographical location, housing type, and other factors on the planning and design of residential areas and the construction of public space, we finally determined 10 samples of residential areas that differ in aspects of the construction time, geographical location, housing type, plot ratio, and greening rate (Figure 2). After field visits, we can confirm that the selected samples can cover the overall construction circumstances of old residential areas in Hefei and are fully typical and representative (

Table 1. Basic information of investigated residential areas.

Names of Residential Areas	Administrative District	Geographical Location	Housing Type	Time of Completion	Volume Fraction	Greening Rate
Xinglin Garden	Luyang	Inside the Second Ring	Housing reform	2000	1.3	0.27
Jinniao Garden	Luyang	Inside the Second Ring	Housing reform	1995	2.5	0.2
Lianquan Residential Area	Shushan	Inside the Second Ring	Housing reform	1989	1.6	0.1
Municipal Government Agency Dormitory	Luyang	Inside the First Ring	Unit collective self-built house	1988	1.2	0.2
Anhui Provincial Construction Department Compound	Luyang	Inside the First Ring	Unit collective self-built house	1991	2	0.2
No. 3 Living area of Anhui No. 1 Construction Company	Shushan	Inside the Second Ring	Unit collective self-built house	1990	1.2	0.2
Emerald Garden	Shushan	Inside the First Ring	Commercial housing residence	2000	2.8	0.4
Jin’an Garden	Baohe	Inside the Second Ring	Commercial housing residence	2000	1.6	0.4
Gangbei No. 2 Village	Yaohai	Outside the Second Ring	Relocation supporting house	1989	2.31	0.3
Feinan Living area	Baohe	Outside the Second Ring	Relocation supporting house	2000	2.1	0.3

). The set of pictures in Figure 3 illustrates the current situation of public space use in the samples, which is also a common phenomenon of public space in old residential areas of Hefei.

2.2. Methods

2.2.1. Impact-Asymmetry Analysis

IAA originates from the three-factor theory. Based on the pioneer work by Kano et al., the three-factor theory categorized factors into three types, namely, excitement factors, performance factors, and basic factors [32]. As can be seen in Figure 4, the influence of excitement factors and basic factors on overall satisfaction is nonlinear, as opposed to the linear influence of performance factors on overall satisfaction. Therefore, the three-factor theory posits that basic factors must be improved, thus enjoying the highest improvement priority, while excitement factors have the lowest improvement priority, since the poor performance of excitement factors can hardly reduce the overall satisfaction. Performance factors are second in improvement priority. While the three-factor theory identifies the priority of improvement and classifies service factors, it fails to consider the influence of factors on overall satisfaction. In 2008, Mikulic and Prebezac proposed IAA [33], which extends penalty–reward contrast analysis (PRCA), evaluates the asymmetric range of influence on satisfaction by factors and uses a value called IA (impact asymmetry) to quantify the asymmetric impact of factors on overall satisfaction. Therefore, IAA classifies the different types of factors in the three-factor theory more finely and can prioritize the evaluated factors more accurately [34]. As shown in Figure 4, the factors are divided into five types [18]. Among them, delighters and satisfiers are equivalent to the excitement factor in the three-factor theory, just as dissatisfiers and frustrators are to the basic factor and hybrids are to the performance factor. The IAA method has been gradually introduced into the field of urban planning in recent years. Cao, Dong, Fang, and others adopted this method to classify neighborhood factors, walking factors, and bus service factors in cities [18,31,34].

2.2.2. Gradient-Boosting Decision Trees

A GBDT is used in this study to assess the relative importance of different environmental factors in public spaces. As an ensemble learning algorithm based on decision trees, it improves prediction performance by training multiple decision tree models step by step [35]. The purpose is to account for the prediction error by iteration until the loss function remains steady or becomes minimized, so that the final result is closer to the true value [31].

When making a comparison between the traditional regression model and GBDTs, the latter can generate predictions of higher precision and eliminate the need for data to conform to a specific distribution; GBDTs are also capable of handling the missing data of variables, which helps tackle the multicollinearity problem [31]. The gradient enhancement of decision trees can result in a strong prediction model. The final output formula of the GBDT model is:

$$\begin{array}{c}{F}_m\left(x\right)=F_{m-1}\left(x\right)+\xi {\displaystyle \sum _{j=1}^J}{\gamma }_{jm}I\left(x\in R_{jm}\right),0<\xi \le 1\end{array}$$

(1)

where ${\gamma }_{jm}$ is the single optimal value for each region $R_{jm}$, when $x\in R_{jm}, I=1$; otherwise, $I=0$. $\xi$ is the contraction parameter, i.e., the learning rate. Regarding a single decision tree $T$, in order to estimate the relative importance of the predictor $x_k$ in the prediction response, the following formula is used:

$$\begin{array}{c}{I}_k^2\left(T\right)={\displaystyle \sum _{t=1}^{J-1}}{\widehat{\tau }}_t^{2}I\left(\nu \left(t\right)=k\right)\end{array}$$

(2)

where $J$-terminal node tree $T$ is the sum of non-terminal nodes $t$, $x_k$ is the splitting variable related to node $t$, and ${\widehat{\tau }}_t^2$ is the reduction in the square error as the result of using the prediction factor $x_k$. For a collection of decision trees ${\left\{T_m\right\}}_1^M$, it can be generalized by its average over all trees in Formula (2):

$$\begin{array}{c}{I}_k^2=\frac{1}{M}{\displaystyle \sum _{m=1}^M}{I}_k^2\left(T_m\right)\end{array}$$

(3)

2.3. Data

2.3.1. Variables

The establishment of an indicator system for public space environment evaluation lays a foundation to form a systematic understanding of the structural texture and functional organization of public space in old residential areas. Since the public space in residential areas is essentially the spatial carrier of the daily needs of residents, a functional system of public space can be deconstructed from the perspective of residents’ needs. This study chose Maslow’s hierarchy of needs theory as the basic framework (Figure 5), and after synthesizing existing research results [12,18,25,26,31,36], a public space environment evaluation system oriented to residents’ needs was formed (

Table 2. Evaluation indicator system of public space environment.

Two Dimensions	First Level Indicators	Secondary Indicators	Tertiary Indicators
Physical environment	Traffic space	Traffic network design	Connectivity of traffic routes
		Pavement quality	Pavement levelness
		Road width	Roadway width
			Sidewalk width
		Parking lot	Motor vehicle parking lot Non-motor vehicle parking lot
		Internal traffic in buildings	Comfort of walking passages in buildings
		Barrier-free traffic	Barrier-free facilities
	Green landscaping space	Plant configuration	Green area Floristics
		Landscape shaping	Vegetative landscape Sketch landscape
	Outdoor recreation space	Desk and chair leisure facilities	Public seat configuration
		Sports and fitness facilities	Configuration of sports and fitness facilities
			Practicability of sports fitness facilities
		Children recreational facilities	Configuration of children recreational facilities Safety of children recreational facilities
		Sound insulation and noise prevention facilities	Effectiveness of sound insulation and noise prevention
	Life service space	Living service facilities	Convenience of commercial facilities Garbage collection point configuration
		Security facilities	Access control management facilities
			Monitoring facilities
			Lighting facilities
			Fire-fighting device
			Tag system
		Public building	Building quality
			Reading and learning space
			Display and communication space
			Nursing space for the old and young
Social environment	Community service	Property service	Property service level
		Management and maintenance	Hygiene and cleanliness
			Safety-check
			Greening maintenance Facility maintenance
		Activity organization	Frequency of holding activities Diversity of activities Information transmission efficiency
	Cultural atmosphere	Cultural propaganda	Cultural display and publicity
		Neighborhood communication	Participation frequency of public activities Degree of communication with neighbors
		Resident behavior	Occupation of public space (parking spaces, sidewalks, staircases, etc.) Uncivilized behavior of residents (spitting, littering, etc.)

). This indicator system includes two dimensions, physical environment and social environment, covering six first-level indicators, namely traffic space, green landscaping space, outdoor recreation space, life service space, community service and cultural atmosphere; 21 secondary indicators; and 42 tertiary indicators corresponding to the seven types of needs, i.e., physiological needs, safety needs, belongingness and love needs, self-esteem needs, cognitive needs, aesthetic needs, and self-actualization needs. Based on this indicator system, the questionnaire design was completed. The questionnaire adopted a 7-level Likert scale, through which respondents can evaluate environmental factors and overall satisfaction with public space on a scale of 1–7 according to their true feelings. The content of the questionnaire incorporates five parts: satisfaction from physical environment factors, satisfaction from social environment factors, the overall satisfaction of residents, basic information about the residents, and the characteristics of residents’ activities. According to the distribution results of the questionnaire, both the indicator system and the questionnaire have been well verified and can conform to the actual needs of residents accurately.

2.3.2. Statistics

The research team conducted face-to-face interviews randomly in the 10 old residential areas, for example, Xinglin Garden, and distributed questionnaires. From November to December 2023, 362 research questionnaires altogether were collected. After the exclusion of 14 from the total as invalid responses, the remaining valid ones were 348, with a validity rate of 96.7%. Among the respondents, the number of females slightly outweighed that of males, mainly because the research work was mostly carried out during the working day when more middle-aged and elderly females tend to stay in the community. Moreover, females exhibited a higher degree of cooperation with the research. Respondents include all ages, mainly middle-aged and elderly people, followed by young respondents. The average education level is slightly lower, with 60% of the respondents receiving a high school education or less. The respondents have enjoyed a long period of residence on average, with 27.0% having lived there for 8 to 15 years and 22.3% for more than 15 years. In

Table 3. Demographics of respondents.

Characteristics	Category	Percentage (%)
Gender	Female	45.7
	Male	54.3
Age	17 and under	4.6
	18–28	17.0
	29–40	23.0
	41–55	23.5
	56–65	23.0
	66 and above	8.9
Education level	Elementary school or lower	14.7
	Middle school	24.1
	High school/vocational high school	19.5
	Junior college	23.0
	Bachelor’s degree/associate degree	15.5
	Graduate degrees	3.2
Occupation	Students	12.1
	Office workers	27.6
	Individual workers	8.3
	Freelancers	17.2
	Retirees	15.5
	Others	19.3
Number of cars owned	0	29.5
	1	58.0
	2	11.7
	3 and above	0.5
Duration of residence (year)	Less than 1	12.6
	1–3	17.2
	3–8	20.6
	8 to 15	27.0
	15–20	8.6
	More than 20	13.7

are the demographic data of respondents.

3. Results

3.1. Analysis of Importance of Environmental Factors in Public Space

When employing the GBDT method in the framework of PRCA, the initial step involves the conversion of the satisfaction evaluation scores of 42 public space environmental factors into dummy variables, with satisfaction scores ranging from 1 to 3 categorized as penalties and recoded as −1, the score of 4 designated as the reference index and recoded as 0, and scores from 5 to 7 considered rewards and recoded as 1 [37]. Subsequently, a PRCA is conducted, taking two pairs of dummy variables recoded previously as independent variables and the overall satisfaction of residents with the public space environment as the response variable. The GBDT model regarding the public space environment of older residential areas in Hefei is then constructed using the R programming language package “gbm2.1.8”. To enhance the accuracy of the model, the learning rate is set at 0.001, and overfitting is mitigated through five-fold cross-validation. The final iteration ordinal number is 3063, with the cross-validation error reaching 0.2981348.

In

Table 4. Relative impact of environmental factors in public space on overall satisfaction.

Serial Number	Indicators	Relative Impact (%)
1	Uncivilized behavior	13.44
2	Space occupation	9.79
3	Nursing space for the old and young	7.08
4	Hygiene and cleanliness	6.53
5	Display and communication space	6.52
6	Sound insulation and noise prevention	5.90
7	Reading and learning space	4.62
8	Building quality	4.18
9	Access control management facilities	4.03
10	Roadway width	3.41
11	Connectivity of traffic routes	3.31
12	Facility maintenance	3.10
13	Monitoring facilities	2.91
14	Cultural display and publicity	2.86
15	Comfort of walking passages in buildings	2.51
16	Information transmission efficiency	2.30

Note: Factors are listed in order of their relative impact on overall satisfaction. Only factors with relative impact values greater than 2% are listed.

are the relative impact values of environmental factors in public space, with values exceeding 2% considered as relatively important factors [31,34,37,38]. Ultimately, the analysis encompassed sixteen public space factors found in old residential areas.

What was initially hypothesized is that among the factors of physical environment in old residential areas, those related to basic renovation, such as parking lots, sidewalk width, pavement quality, and sports and fitness facilities, would significantly affect the overall satisfaction of residents. The results of the GBDT analysis, however, indicate that the subjective perception of residents living in old residential areas is more biased towards resident behavior (space occupation and uncivilized behavior) and public buildings (building quality, reading and learning space, display and communication space, and nursing space for the old and young). Regarding the most common problem of car parking in old residential areas, the average satisfaction with motor vehicle parking lots ranks lower, while the relative impact is very small. As there is not enough space to set up parking lots in old residential areas, problems arise, such as parking lot grabbing and occupation, sidewalk occupation, lawn occupation, and so on. Although residents claim that they are not satisfied with the lack of parking lots, most of them indicate that “the old residential areas cannot help with it”. In other words, residents show their understanding towards the lack of parking lots in the physical environment, but they are dissatisfied with space occupation attached to social environment factors, which exerts stronger effects on overall satisfaction. Satisfaction with the social environment can therefore be enhanced through both management and maintenance to improve the degree to which residents are generally content with public spaces within their places of residence, when it is hard to improve the physical space environment in old residential areas.

3.2. Priorities for Improving Public Space Environment

When environmental factors were expressed as dissatisfied, neutral, and satisfied, the GBDT model was used to calculate the predicted overall satisfaction scores (POSSs), which were expressed as possd, possn, and posss, respectively. The range of the impact on overall satisfaction (RIOS), satisfaction-generating potential (SGP), dissatisfaction-generating potential (DGP), and the impact asymmetry index (IA index) can be calculated based on the POSS results. The calculation formula is as follows:

RIOS = posss − possd

SGP = (posss − possn)/RIOS

DGP = (posss − possd)/RIOS

IA index = SGP − DGP = (posss + possd − 2possn)/(possn − possd)

Different threshold settings in IAA in terms of attribute classification may lead to different factor structures. Recent studies determined 0.2 and −0.2 as thresholds to differentiate linear effects from nonlinear ones [18,30,31,35]. Based on the threshold used by Lee et al. [37,38,39], the present study finally defines attribute categories as follows: delighters (IA index ≥ 0.7), satisfiers (0.7 > IA index > 0.2), hybrids (0.2 ≥ IA index ≥ −0.2), dissatisfiers (−0.2 > IA index > −0.7), and frustrators (IA index ≤ −0.7), which are shown in

Table 5. Impact-asymmetry analysis of environmental factors in public space.

Public Space Environmental Factors	SGP	DGP	RIOS	IA Index	Attribute Type	Average Satisfaction
Uncivilized behavior	0.14	0.86	0.32	−0.72	Frustrator	3.49
Nursing space for the old and young	0.02	0.98	0.25	−0.96	Frustrator	3.02
Space occupation	0.30	0.70	0.24	−0.40	Dissatisfier	3.36
Sound insulation and noise prevention	0.96	0.04	0.23	0.93	Delighter	4.17
Hygiene and cleanliness	0.84	0.16	0.22	0.68	Satisfier	4.06
Building quality	0.38	0.62	0.18	−0.24	Dissatisfier	3.92
Reading and learning space	0.20	0.80	0.17	−0.60	Frustrator	3.01
Connectivity of traffic routes	0.64	0.36	0.15	0.28	Satisfier	4.70
Access control management facilities	0.07	0.93	0.15	−0.85	Frustrator	3.61
Roadway width	0.44	0.56	0.13	−0.12	Hybrid	3.99
Display and communication space	0.04	0.96	0.12	−0.93	Dissatisfier	3.17
Comfort of walking passages in buildings	0.04	0.96	0.11	−0.92	Frustrator	4.08
Monitoring facilities	0.95	0.05	0.11	0.91	Delighter	4.33
Facility maintenance	0.81	0.19	0.10	0.62	Satisfier	3.71
Information transmission efficiency	0.16	0.84	0.08	−0.69	Dissatisfier	4.04
Cultural display and publicity	0.08	0.92	0.08	−0.84	Frustrator	3.39

. In view of the IA index, most environmental factors of public space share a nonlinear influence on the overall satisfaction of residents, and only the factor of roadway width has a linear relationship.

According to the previous research on IAA [37,38,39], the priorities for improvement should comprehensively consider the attribute, influence scope, and performance of the factors. The average satisfaction value of 42 factors covering the public space environment is used as a reference point to measure the performance of each factor at present. In this paper, the average value of overall satisfaction for public space environment factors in old residential areas is 3.80. As is presented in Figure 6, the IA index and RIOS are used to draw an IAA diagram of environmental factors for the public space of old residential areas in Hefei. To compare the relative sizes of the impact ranges of various factors more intuitively, RIOS is divided into high, medium, and low impact ranges, with the formula expressed as follows:

(1)

The range of high impact: RIOS > (Max [RIOS] + Avg [RIOS])/2;

(2)

The range of medium impact: (Min [RIOS] + Avg [RIOS])/2 < RIOS < (Max [RIOS] + Avg [RIOS])/2;

(3)

The range of low impact: RIOS < (Min [RIOS] + Avg [RIOS])/2.

Figure 6. Impact-asymmetry analysis diagram of environmental factors of public space in old residential areas. Note: The number in parentheses indicates the average satisfaction with that factor.

Figure 6. Impact-asymmetry analysis diagram of environmental factors of public space in old residential areas. Note: The number in parentheses indicates the average satisfaction with that factor.

In the old residential areas of Hefei are two environmental factors that generate a high impact on the overall satisfaction with public space, namely uncivilized behavior and nursing space for the old and young, and the performance of these two is poor, lower than the overall average satisfaction, making them the first to be improved in the old residential areas. Moreover, uncivilized behavior and nursing space for the old and young are both frustrators, an extreme level of dissatisfiers, and when their performance is poor, residents will feel greatly dissatisfied. Therefore, these two factors should be improved first, and resident satisfaction can be greatly improved, provided that they are to meet residents’ basic needs.

A moderate impact is produced by eight factors of overall satisfaction with the public space environment. Access control and management facilities are frustrators, an extreme level of dissatisfiers. Their average satisfaction is lower than the average value of overall satisfaction, which will make residents feel very dissatisfied when performing poorly. Therefore, this factor should be improved first and should merely satisfy the basic needs of residents so as to optimize investment. Building quality, space occupation, and reading and learning space are dissatisfiers, among which space occupation and reading and learning space perform poorly. Improving them to satisfy the basic needs of residents can greatly improve the overall satisfaction of residents, so these two factors should be considered first. Although building quality is a dissatisfier, its average satisfaction is somewhat higher than the average value of overall satisfaction, and its performance is slightly better. As its effect on satisfaction is not as good as the dissatisfiers with poor performance, it should be improved after the frustrators and other dissatisfiers that have a moderate impact. Sound insulation and noise prevention are delighters, an extreme level of satisfiers. When their performance is poor, overall satisfaction will not be greatly reduced. Provided their performance exceeds the expected level of residents, the overall satisfaction of residents will be significantly improved. The average satisfaction with sound insulation and noise prevention exceeds the average value of overall satisfaction, and its performance is better. Therefore, it should be improved after the under-performing dissatisfiers and frustrators, which can greatly improve the overall satisfaction of residents. The connectivity of traffic routes and hygiene and cleanliness are satisfiers, which are similar to delighters. However, their average satisfaction exceeds the overall average satisfaction, making their effect on overall satisfaction not as strong as that of delighters; thus, their improvement should come after delighters. Roadway width is a hybrid with good performance, which can come after other significant factors in terms of improvement.

Six factors are found to produce a low impact regarding overall satisfaction with the public space environment. Cultural displays and publicity, the comfort of walking passages in buildings, and display and communication space are frustrators, an extreme level of dissatisfiers. Poor performance of these factors will make residents feel very dissatisfied. Among them, the average satisfaction with cultural display and publicity as well as display and communication space are lower, but as they exert little impact on the overall satisfaction, both factors should follow the first ones with respect to improvement. For the sake of investment optimization, these two are to be improved to the extent of meeting residents’ basic needs. The average satisfaction level with the comfort of walking passages in buildings reaches 4.08, which is higher than the overall average satisfaction degree and has reached the level of meeting the basic needs of residents. Further improvement will not greatly enhance the overall satisfaction degree; thus, it should be improved after cultural displays and publicity and also display and communication space. Information transmission efficiency is a dissatisfier, which is similar to a frustrator. When its performance is not good, the effect after being improved is not as strong as that of the frustrators on overall satisfaction, and its average satisfaction exceeds the overall average satisfaction. Therefore, its improvement order should be placed after the frustrators. Monitoring facilities are delighters, the extreme level of satisfiers. When the performance is poor, the overall satisfaction will not be greatly reduced. When the performance exceeds the expected level of residents, the overall satisfaction of residents will be significantly improved. The average satisfaction with monitoring facilities exceeds the average satisfaction, and their performance is good. Therefore, they should be improved after the dissatisfiers with poor performance and frustration factors, which can greatly improve the overall satisfaction of residents. Facility maintenance is a satisfier, which is similar to the delighters. Because its average satisfaction is lower than the overall average satisfaction, the effect gained through improvement is not as strong as the delighters, so its improvement order should be placed after the delighters. However, since monitoring facilities and facility maintenance have a low impact on overall satisfaction, improving these two elements cannot significantly improve resident satisfaction. Therefore, they should be improved after the optimization of more important environmental factors.

4. Discussion

China is currently undergoing a profound transformation from large-scale urbanization and incremental urban construction to urban renewal to promote urban sustainable development. This study focuses on renovation practices in public space in old residential areas and continues to explore scientific issues in this field. The deconstruction and quantification of social environment factors are still limited in the previous research results on the public space environment evaluation of old residential areas in China. In the process of field surveys and interviews regarding old residential areas in Hefei, social environment factors that significantly affect resident satisfaction, such as space occupation and uncivilized behavior, have been noticed. Although these factors are common in some old residential areas, they are rarely mentioned and quantified in existing research. Quantifying these factors and bringing them together into the indicator system of environmental factors of public space in old residential areas can enrich planners’ understanding of other potential factors besides physical ones. Meanwhile, the research results also indicate that uncivilized behavior and space occupation exert a great influence on resident satisfaction, and improving these factors to meet the basic needs of residents can greatly enhance resident satisfaction. This result provides a new reference for future researchers to construct an evaluation system for the environment in residential areas, and it also helps to deepen their understanding of the complex mechanism in terms of the impact of public space on satisfaction in old residential areas. Secondly, this study relaxes the assumption pertaining to the linear relationship between environmental factors and satisfaction, as is widely used in the literature [28], and provides supporting evidence for widespread nonlinear relationships between environmental factors and satisfaction, which further enriches the existing results in the field of nonlinear research. It is worth noting that although Cao et al. and this study both found the universality of nonlinearity, the nonlinear model of specific environmental factors exhibits local heterogeneity due to distinctive emphases attached to the research area and design. Compared with previous research on nonlinear relationships, empirical research on Hefei has made new progress. Parallel with the research results of Song et al. regarding newly built centralized public space in communities in Suzhou [12], this study also found that the social environment attribute is more important than the physical environment attribute among the environmental factors affecting satisfaction with public space. However, this study distinguished more microscopic factors in IAA, and the impact of resident behavior on satisfaction was considered. Additionally, previous studies usually relied on the coefficient of environmental indicators to determine the relative importance of environmental factors, while this study improved the importance hierarchy of environmental factors according to their relative importance and nonlinear correlation with resident satisfaction.

In terms of method application, an integration of the machine learning (GBDT) method into IAA can not only classify the environmental factors that affect resident satisfaction more accurately and determine the improvement priority according to the influence scope and satisfaction score but also predict the small sample size more reliably, which makes up for the problem of the small sample size in this study and significantly reduces the difficulty and time of future research. Cao et al. verified that a sample size of 320 can also provide accurate predictions [30]. In view of practical applications, the existing research results have not yet formed a direct and specific guiding role for the renovation of public space in old residential areas of China. For example, when studying old communities in Harbin, Dong et al. categorized a list of 63 potential contributing factors affecting neighborhood satisfaction into eight groups using factor analysis, after which the factor structure was determined with IAA [18]. However, the meanings of the important influencing factors identified by Cao et al. are still relatively broad, such as housing and public facilities, neighborhood facilities, and the social environment. To clarify the renovation priorities of public space planning intervention, this study, based on the analysis and classification of 42 tertiary indicators, finally determined a factor structure consisting of 16 environmental factors that affect resident satisfaction. The research results can help planners ascertain the priority and goal of public space environmental renovation in old residential areas, so as to enhance residents’ happiness by improving the public space environment in old residential areas and ultimately promote the sustainable development of cities.

It should be further noted that the priority hierarchy of environmental renovation, which was obtained through the IAA method in this study, still contains subjective factors, including the extent of improvements to meet the basic needs of residents, and future research might explore how the basic needs of residents can be quantified by means of objective environmental indicators. Further research might further classify the old residential areas with objective environmental data, explore the similarities and differences pertaining to the factor structures of different types of old residential areas, and quantify the renovation goals that effectively improve resident satisfaction, so as to put forward more detailed and objective suggestions for the renovation of old residential areas and offer much more accurate instructions on the environmental renovation of old residential areas.

5. Conclusions

Based on Maslow’s hierarchy of needs theory, the study established a research framework and constructed an evaluation system for public space environments in old residential areas, with traffic space, green landscaping space, outdoor recreation space, life service space, community service, and cultural atmosphere designated as the first-level indicators from the two dimensions of the physical environment and the social environment. Furthermore, it evaluated satisfaction with the public space environment in old residential areas of Hefei by using IAA and GBDT, and it determined the priority of environmental factors for the public space renovation of old residential areas, guided by the improvement of resident satisfaction. The specific conclusions are as follows:

(1)

Uncivilized behavior and other social environment factors exert the greatest impact on residents’ overall satisfaction. This finding does not support the original belief that the critical problem in the renovation of old residential areas lies in the poor physical environment. According to the research results of satisfaction evaluation and IAA of various factors, residents have a high level of acceptance of the fact that parking lots, roadway width, green areas, and other physical environment factors are difficult to be improved due to space constraints, which poses a lesser impact on overall satisfaction. Therefore, factors of the social environment outweigh those of the physical environment in the public space renovation of old residential areas in Hefei at the current stage.

(2)

There exists a nonlinear relationship between the environmental factors of public space in most old residential areas and the overall satisfaction of residents, which conforms to the conclusion of previous studies. This study further shows that among the environmental factors that generate a high impact on satisfaction, social environment factors are all nonlinear, while the physical environment factors that are supposed to be improved are also the space carrier for creating a fair social environment, such as nursing space for the old and young, reading and learning space, and display and communication space.

(3)

Based on the principle of satisfying the needs of residents and promoting satisfaction, public space environment factors to be improved in old residential areas of Hefei were prioritized. The results showed that resident behavior (uncivilized behavior, space occupation), public buildings (nursing space for the old and young, reading and learning space), and security facilities (access control and management facilities) should be given priority, and resident satisfaction can be greatly improved by simply meeting the basic needs of residents. Secondly, sound insulation and noise prevention, hygiene and cleanliness, the connectivity of traffic routes, cultural displays and publicity, and display and communication space should be improved. Among them, sound insulation and noise prevention, hygiene and cleanliness, and the connectivity of traffic routes are highly satisfactory, which will significantly increase the overall satisfaction of residents after they are improved beyond residents’ expectations. The improvement of cultural displays and publicity as well as display and communication space to meet the basic needs of residents can greatly improve resident satisfaction. Finally, roadway width, monitoring facilities, facility maintenance, building quality, the comfort of walking passages in buildings, and information transmission efficiency should be considered.

(4)

The improvement priority regarding environmental factors basically conforms to Maslow’s hierarchy of needs theory. When residents basically meet their physiological needs and safety needs with space as the carrier, they will place more emphasis on social needs, self-esteem needs, and self-actualization, such as display and communication space for social needs, resident behavior for self-esteem needs, and reading and learning space for self-actualization. Meanwhile, residents will also put forward higher requirements for environmental quality, such as sound insulation, noise prevention, hygiene, and cleanliness. Therefore, for those old residential areas that have significant space constraints and have undergone the first round of renovation, the subsequent environmental renovation work of old residential areas in Hefei should pay more attention to the social needs, self-esteem needs, and self-actualization needs of residents; restrain resident behavior through planning and governance; and improve the level of property service. At the level of the physical environment, both old and young residents should be provided with space for activities, display and communication, and reading and learning through renewal and new construction.