CRITIC-TOPSIS Based Evaluation of Smart Community Governance: A Case Study in China

Abstract

As the basic unit of a smart city, the smart community has received considerable attention and problems in community governance have appeared simultaneously. Previous studies of smart community governance have failed to encompass all aspects, especially the evaluation tools for ensuring its outcomes. Therefore, this paper developed a comprehensive evaluation framework based on the CRITIC-TOPSIS method combined with the identified evaluation indicators. Seven smart communities from four cities in China were selected as cases to show how this evaluation framework could be applied to decision-making. The results indicated that the evaluation indicator ”Mediation of Conflict” had the highest weight while ”The participation of social enterprises in governance” had the lowest weight. Furthermore, the Yucun community presented the highest governance performance among these seven smart communities. Several strategies are proposed for improving the level of smart community governance, such as devoting significant resources to develop infrastructure in smart communities, facilitating communication among multiple participants, and increasing funding for the implementation of smart communities. This research contributes both to the innovation of community governance evaluation and to the improvement of smart communities.

1. Introduction

With the development of urbanization, the smart community, which is the fundamental unit of smart cities [1,2,3,4,5,6,7], has been seen as a promising solution to urban problems [8,9,10]. Generally, a smart community is a community that applies the Internet of Things (IoTs), cloud computing, big data, and other new information technologies to digitize and coordinate community residents’ daily lives [3,11,12]. In the development of communities, there are a series of problems hindering the implementation of the smart community, such as a lack of unified planning and deployment, a serious “information island” and repeated construction [13]. Especially with regard to community governance, the issues of mismatching demands, a shortage of funds, and insufficient residents’ participation bring challenges to smart community construction [9,14]. To address these problems, smart community governance has attracted increasing attention and it contributes to improving the system and capacity of national governance [15]. In general, smart community governance involves the collaborative management of public affairs and the delivery of community services by using intelligent tools [16,17,18]. In this process, it is necessary to promote the participation of stakeholders, such as government departments, public welfare departments, enterprises, and residents [15,16,17,18]. Moreover, smart community governance consists of three critical elements: addressing community needs, maintaining public order in smart communities, and promoting the long-term development of such communities [17,18,19]. In China, smart community construction is also affected by community governance issues [14,20,21]. For example, due to the lack of a platform for communication, few residents in the community of Suzhou city participated in community affairs, which affects the improvement of community governance in such a community [22]. As a useful tool for ensuring the outcomes of smart community governance, performance evaluation could be taken for improving stakeholders’ strategies or solutions [23,24,25]. Specifically, by conducting the evaluation of smart community governance, more effective funding arrangements could be made, and the capability of community service and management could be increased [20,26]. Hence, it is necessary to establish a comprehensive and feasible evaluation system for smart community governance.

The existing research has made enormous contributions to the development of smart community governance. For example, technical systems such as the smart community integrated multi-generation energy system [27], smart community information security encryption scheme [28], and smart community governance system [26] are designed to assist smart community governance. Meanwhile, empirical progress has been made in the evaluation of community livability through building the evaluation indicator system [29]. An evaluation index system for transportation in smart communities has been designed to deal with the transportation problems that often occur in the smart community [30]. Furthermore, an indicator system was developed to evaluate the level of community governance from the perspective of participatory governance, and 33 secondary evaluation indicators were included in this system [31]. A prospect theory-based evidential reasoning (PTER) analytical framework is proposed to comprehensively evaluate the smart community for sustainable development [20]. In addition, in order to achieve continuous improvements in service quality, a contrast approach to benchmarking smart community health centers is developed [32]. However, most studies have focused solely on one aspect of smart community governance, and the comprehensive evaluation indicators specifically for assessing smart community governance are currently unavailable. Furthermore, most of the evaluation methods used are subjective and based on expert judgments or public questionnaires. A quantitative evaluation of smart community governance is needed in combination with the actual situation of smart communities.

Performance evaluation of smart community governance plays a crucial role in ensuring the effectiveness of its implementation and development. Therefore, assessing the level of smart community governance is vital for providing new insights into the development of community governance and promoting residents’ engagement. Thus, considering the gap in existing studies, this paper aims: (1) to identify the performance indicators of smart community governance; (2) to establish a framework for the performance evaluation of smart community governance through the CRITIC-TOPSIS method, and (3) to provide suggestions that will help improve the long-term governance of smart communities.

2. Methodology

In order to quantify the level of the smart community governance, an evaluation indicator system of the smart community governance was established, and the evaluation model was developed through the CRITIC (Criteria Importance Through Intercriteria Correlation)-TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) method. The evaluation process of smart community governance is given in the following steps (as illustrated in Figure 1).

Step 1. Identifying the preliminary evaluation indicator system of the smart community governance through a systematic literature review (SLR).

Step 2. Selecting the evaluation indicators for smart community governance through expert interviews.

Step 3. Evaluating the performance of smart community governance construction through the CRITIC-TOPSIS method.

2.1. Identifying the Evaluation Criteria for Smart Community Governance

Extracting evaluation indicators is the basis for evaluating the governance performance of smart communities. The evaluation indicators of smart community governance are extracted directly from articles and guidelines. Based on the SLR method (as shown in Figure 2), the 30 studies went through an in-depth confirmation process to ensure that they met the research objectives. Furthermore, evaluation guidelines related to smart communities and community governance were referenced for the evaluation indicator system. Two distinct types of evaluation indicators were found during the indicator extraction process: (1) Some indicators implement similar purposes or even the same purpose while their names are different from each other; (2) Some indicators with the same names for smart community governance are supported by previous articles or guidelines. Using all collected indicators in the following analysis would be unreasonable. Thus, further selection of indicators is essential.

For picking out eligible and reliable indicators, the frequency of these indicators mentioned in studies is chosen to apply to the indicator identification. Generally, a more frequent evaluation indicator suggests that this aspect is crucial to smart community governance because more academics believe it merits further investigation. Consequently, the rules of indicator identification are set as follows: (1) Check whether the notion of indicators is similar or complementary. If any, these indicators should be combined in the same dimension; otherwise, these indicators should be divided into different dimensions; (2) The indicator which has been mentioned more than three times should be selected; otherwise, this indicator should be removed; (3) Particularly, a few indicators that have been mentioned only once but are of great significance for the evaluation should also be selected [33]. The preliminary evaluation indicator system was then determined with six dimension indicators and 25 second-level indicators, as shown in

Table 1. Preliminary evaluation indicator system of smart community governance.

Dimension	Indicators	Sources
Multiple Participation	Grid-based Governance, Resident Autonomy, Participation of Social Organizations Governance	[14,34,35,36,37,38,39,40,41,42,43,44,45]
Object Management	Population Management, Management of Community Organizations, Party Building Management, Volunteer Management, Vehicle Management, Housing Management	[35,36,46,47,48,49]
Government Services	Information Disclosure, Administrative Examination and Approval, Services for the Elderly, Services for Economically Disadvantaged Persons, Services for Persons with Disabilities, Services for the Floating Population	[14,15,35,36,39,50,51,52,53,54,55,56]
Public Safety Management	Key Regional Management, Special Population Management, Police Citizen Interaction, 110 Alarm Service	[35,36,37,47,57,58,59,60,61,62]
Adjustments and Corrections	Legal publicity, Mediation of Conflict, Legal Services	[35,46,63,64]
Community Culture	Community Activities, Recreation and Sports, Community Communication	[46,48,65,66]

.

2.2. Selecting the Evaluation Indicators for Smart Community Governance

The evaluation indicator for smart community governance is mainly obtained from literature and the guidelines for smart communities and community governance; it is necessary to optimize the evaluation indicators to make them more relevant to better reflect the status of smart community governance. Therefore, 53 scholars in the field of smart community research and experts in the practice field were interviewed online by a VooV Meeting in June 2022. The relevant information of the experts is shown in

Table 2. Profiles of valid interview respondents.

Individual Characteristics	Items	Percentage (%)
Gender	Male	58.49%
	Female	41.51%
Education level	Doctorate degree	18.81%
	Master’s degree	50.94%
	Other	30.19%
Working experience	More than 5 years	18.87%
	3 to 5 years	16.98%
	1 to 3 years	26.42%
	Less than 1 year	37.74%
Occupation	College teacher	18.87%
	Government employee	5.66%
	Manager of an enterprise	45.28%
	Other	30.19%

. The importance and content of the preliminary evaluation indicator system were adjusted through expert interviews. The indicator would be accepted if the proportion of experts who consider the indicator to be very important and important exceeds 80% [33,67]. In the evaluation indicator system, “Participation of social entrepreneurs in government“ was added while the indicators “110 Alarm Service“ and “Recreation and Sports“ were removed. The final evaluation indicator system was established, which includes six dimension indicators and 23 second-level indicators, as shown in

Table 3. Final evaluation indicator system of smart community governance.

Dimension	Indicators	Code
Multiple Participation B1	Grid-based Governance	C11
	Resident Autonomy	C12
	Participation of Social Organizations in Governance	C13
	Participation of Social Enterprises in Governance	C14
Object Management B2	Population Management	C21
	Vehicle Management	C22
	Party Building Management	C23
	Volunteer Management	C24
	Housing Management	C25
Government Services B3	Information Disclosure	C31
	Administrative Examination and Approval	C32
	Services for the Elderly	C33
	Services for Economically Disadvantaged Persons	C34
	Services for Persons with Disabilities	C35
	Services for the Floating Population	C36
Public Safety Management B4	Key Regional Management	C41
	Special Population Management	C42
	Police Citizen Interaction	C43
Adjustments and Corrections B5	Legal Publicity	C51
	Mediation of Conflict	C52
	Legal Services	C53
Community Culture B6	Community Activities	C61
	Community Communication	C62

Note: Detailed information about the evaluation indicator system of smart community governance is shown in Supplementary Materials File S2.

.

2.3. Evaluating the Performance of Smart Community Governance Construction

This paper uses the CRITIC-TOPSIS model to calculate the composite score of smart communities under assessment indicators to evaluate and rank the governance performance of the smart communities. Generally, the CRITIC method determines weights based on the intensity of contrast and conflict evaluation of the decision-making problem [68]. It is developed to avoid the nature of subjectivity, which makes prioritizing one criterion over the others a complex task when multiple responses are involved [69,70]. Hence, the CRITIC method was selected to calculate the weight of each indicator in order to avoid the conceptual synchronization of indicators affecting the result in this study. On this basis, a method that can calculate the performance of smart community governance by using the weight of these indicators is required. Considering that the TOPSIS method makes full use of attribute information to provide decision-makers with a ranking of alternatives [71], and does not require attribute preferences to be independent compared to other multiple criteria models [72,73], it was further decided to offer a cardinal ranking of smart community governance. The procedures for an evaluation model construction for smart community governance by using the CRITIC-TOPSIS method are summarized as follows.

(1)

Calculating weight values between indicators through the CRITIC method.

(a)

The decision matrix is normalized.

The data are normalized in order for their units to be uniform, and their values are between [0, 1]. In this study, all indicators are positive indicators. Positive indicators are preprocessed using Equation (1)

$$x_{ij}^{’}=\frac{x_{ij}{-x}_{min}}{x_{max}{-x}_{min}}$$

(1)

where $x_{ij}$ is the score of the i$-th$ expert in the j$-th$ indicator for the evaluation of smart community governance, ${x\text{’}}_{ij}$ is the value of $x_{ij}$ after normalization, $x_{max}$ is the maximum score of n$-th$ experts in j$-th$ indicator for the evaluation of smart community governance, and $x_{min}$ is the minimum score of n$-th$ experts in j-th indicator for the evaluation of smart community governance.

• (b)

  The standard deviation of each criterion is calculated.

$$\overline{x_j}=\frac{1}{n}{{\displaystyle \sum }}_{i=1}^n{x\text{’}}_{ij}$$

(2)

$$S_j=\sqrt{\frac{{\sum }_{p=1}^n{{(x}_{pj}^{’}-\overline{x_j})}^2}{n}}$$

(3)

$\overline{x_j}$ is the mean score of the $j-th$ indicator for the evaluation of smart community governance. $S_j$ is the standard deviation of the $j-th$ indicator for the evaluation of smart community governance. n is the total number of experts and m is the total number of indicators. ${x\text{’}}_{pj}$ is the normalization value of the p$-th$ expert in the j$-th$ indicator for the evaluation of smart community governance.

• (c)

  Correlation coefficients are calculated.

$$r_{ij}=\frac{{\sum }_{p=1}^n{(x}_{pi}-\overline{x_i}{)(x}_{pi}-\overline{x_j})}{\sqrt{{\sum }_{p=1}^n{{(x}_{pi}-\overline{x_i})}^2\sum {{(x}_{pj}-\overline{x_j})}^2}}$$

(4)

$r_{ij}$ is the correlation coefficient of the i$-th$ indicator and the $j-th$ indicator. $x_{pi}$ is the score of the p$-th$ expert in the i$-th$ indicator for the evaluation of smart community governance. $x_{pj}$ is the score of the p$-th$ expert in the j$-th$ indicator for the evaluation of smart community governance. $\overline{x_i}$ is the mean score of $i-th$ indicator for the evaluation of smart community governance.

• (d)

  The conflict of indicators is calculated.

$$R_j={\displaystyle \sum }_{j=1}^n{(1-r}_{ij})$$

(5)

$R_j$ indicates the conflict of the indicators for the evaluation of smart community governance.

• (e)

  The information content is calculated.

$$C_j{=S}_j{\displaystyle \sum }_{j=1}^n{(1-r}_{ij}{)=S}_j{\text{×}R}_j$$

(6)

where $C_j$ denotes the information content of $r_{ij}$.

• (f)

  The objective weight is determined.

$$W_j=\frac{C_j}{{\sum }_{j=1}^m{C}_j}$$

(7)

$W_j$ is the weight of the $j-th$ indicator for the evaluation of smart community governance.

(2)

Evaluating smart community governance through the TOPSIS method.

The weight of the evaluation indicator (i.e.,$W_j$) is assigned to the TOPSIS model matrix after it is obtained. The procedures for applying the TOPSIS method are described as follows.

• (a)

  Normalization for the value of smart community.

The process of normalization was addressed in the above Equation (1).

• (b)

  Calculate weight values of the normalized score of smart community.

$${\lambda }_{tj}{=W}_j{\theta }_{tj}$$

(8)

where ${\lambda }_{tj}$ is the weight value for normalized indications for the evaluation of smart community governance, ${\theta }_{tj}$ is the value of the t$-th$ smart community in the j$-th$ indicator for the evaluation of smart community governance.

• (c)

  Determine the positive ideal solution ($A^{+}$) and the negative ideal solution ($A^{-}$)

The ideal point ($A^{+}$) is a composite of the best performance values of a sample smart community governance across all indicators, while the negative ideal point ($A^{-}$) is a composite of the worst performance values.

$$A^{+}=\left\{{\lambda }_1{}^{+}{,\lambda }_2{}^{+}\cdots {,\lambda }_m{}^{+}\right\}$$

(9)

$$A^{-}=\left\{{\lambda }_1^{-}{,\lambda }_2^{-}\cdots {,\lambda }_m^{-}\right\}$$

(10)

where

$${\lambda }_j{}^{+}{=max}_t\left\{{\lambda }_{tj}\right\} j=1,2,\cdots ,m,$$

(11)

$${\lambda }_j^{-}=min\left\{{\lambda }_{ij}\right\} j=1,2,\cdots ,m,$$

(12)

${\lambda }_j{}^{+}$ is the maximum weight value for the normalized score of $j-th$ indicator among sample smart communities. ${\lambda }_j{}^{-}$ is the minimum weight value for the normalized score of $j-th$ indicator among sample smart communities.

• (d)

  Calculate distance measures between sample smart communities and the best condition ($D^{+}$) and worst condition ($D^{-}$).

$$D_j{}^{+}=\sqrt{{\sum }_{j=1}^m{{(\lambda }_{ij}{}^{+}{-\lambda }_{ij})}^2}, j=1,2,\cdots ,t,$$

(13)

$$D_j^{-}=\sqrt{{{\displaystyle \sum }}_{j=1}^m{{(\lambda }_{ij}{-\lambda }_{ij}^{-})}^2}$$

(14)

• (e)

  Determine the closeness coefficient.

The value of the closeness coefficient ($T_j$) is used to indicate the relative closeness of a particular sample smart community $t$ to the negative ideal point. A larger value of closeness indicates better performance of the smart community governance [72]:

$$T_j=\frac{D_j^{-}}{D_j^{+}{+D}_j^{-}}$$

(15)

The final score and rankings of the smart community are shown through the TOPSIS method. $T_{B1}$,$T_{B2}, T_{B3}{, T}_{B4}{, T}_{B5}$, and $T_{B6}$ are used to represent the performance value in six dimensions. In addition, the total performance $T_{total}$ for the smart communities is calculated by Equation (16):

$$T_{total}{=T}_{B1}{+T}_{B2}{+T}_{B3}{+T}_{B4}{+T}_{B5}{+T}_{B6}$$

(16)

3. Case Study

3.1. Selection of Smart Communities in China

Four cities in China, namely Beijing, Shanghai, Guangzhou, and Shenzhen, were selected as investigation cities for the following reasons. On the one hand, these four cities have the stronger comprehensive strength among all cities of China. The data of smart community governance are easily collected, which makes it possible to obtain the evaluation data. On the other hand, through the smart community developed earlier in these four cities, extensive experience has been accumulated. Therefore, seven typical communities in these four cities were selected as sample collection sites, including one smart community in Guangzhou (N1: Moxi community), one smart community in Shenzhen (N2: Yucun community), two smart communities in Shanghai (N3: Songshan community and N4: Maqiao community), and three smart communities in Beijing (N5: Huamao Center community, N6: Jinbangyuan community, and N7: Tiantong Dongyuan 2nd community).The areas of the N1–N7 smart communities are approximately 0.30 km², 0.25 km², 0.20 km², 0.43 km², 0.50 km², 0.20 km² and 0.35 km² respectively. The typical feature of these seven communities is that they have made substantial progress in the construction of smart communities. Specifically, the integrated platform for community management has been developed in both N1 and N2. Such platforms realize intelligence in managing community affairs, particularly in the provision of community services with remarkable accuracy and efficiency. In addition, the N3 and N4 were selected as pilot smart communities in Shanghai, and Internet of Things technology was applied to collect the health data of residents in these communities. Similarly, the N5, N6, and N7 have also been chosen as pilot smart communities in Beijing, and smart infrastructures have been developed to cope with emergencies such as epidemics. Hence, these seven communities are selected as typical cases for the evaluation of smart community governance. Figure 3 shows the locations of the selected smart communities.

3.2. Data Collection

The weight data of governance evaluation indicators of smart communities were collected through an expert interview which was conducted online via VooV Meetings in June and July 2022. Each interview lasted approximately one hour and was recorded. The experts were invited to score the important level of evaluation indicators for smart community performance from 1 (very unimportant) to 5 (very important), and the final scoring data were used to calculate the weight of each evaluation indicator. The detailed information about the interview outline can be seen in Supplementary Materials File S1, and the information about the interviewed experts is shown in Table 2. The reliability value of the importance level of 23 second-level indicators (indicated by Cronbach α) was 0.896, suggesting that the data had satisfactory reliability. All the experts have engaged in the implementation of smart communities and community engagement, which means that these respondents have extensive professional experience in smart community governance. Hence, the weight data scored by the experts can be further analyzed and provide the basis for determining the weights by the CRITIC method.

In addition, the values of evaluation indicators for smart community governance were obtained according to the objective criteria. Following these criteria, data for seven communities were collected directly from official websites, WeChat public accounts, and news reports. Meanwhile, supporting materials about seven smart communities were obtained through interviews with community administrators. Corresponding indicator values of seven smart communities were obtained, which provided a basis for the TOPSIS method to calculate smart community governance scores.

4. Result

4.1. Results of Indicators Weight through the CRITIC Method

Based on extant literature and expert opinions, the criteria for the values of evaluation indicators were established. The objective data and supporting materials of smart communities are required to ensure that the value data of evaluation indicators are not affected by subjective factors of data collectors. Details of the criteria for each evaluation indicator are shown in Supplementary Materials File S3. Experts were interviewed to score each evaluation indicator, and the weight of evaluation indicators was calculated through the CRITIC method. The weight results of indicators were obtained from experts’ opinions through the CRITIC method. Experts were interviewed to score each indicator to calculate the weight of each evaluation indicator. The weight of evaluation indicators is shown in Figure 4. According to the result, the three highest weights were the C52, the C33, and the C22; this means that the C52, the C33, and the C22 were the priority for smart community governance. Meanwhile, the three lowest were the C14, the C43, and the C25. The weights of indicators range from 0.0265 to 0.0672, and the average weight of the indicators is 0.0435. There are two indicators with a weight between 0.02 and 0.03, accounting for 8.7%. The weights of four indicators are between 0.03 and 0.04, accounting for 17.4%. The number of indicators with a weight between 0.04 and 0.05 is the largest, with 14 indicators, accounting for 60.9%. There is one indicator between 0.05 and 0.06 and two indicators above 0.06, accounting for 4.3% and 8.7%, respectively. The weights of indicators provided the basis for the evaluation of smart community governance.

4.2. Results of Evaluation through the TOPSIS Method

After obtaining the weight for each indicator, the performances of sample smart communities were calculated through the TOPSIS method. The performance results are illustrated in

Table 4. Assessment results of smart community governance from the TOPSIS method.

	Community	N1	N2	N3	N4	N5	N6	N7
Dimension
Multiple participation (B1)		0.316	0.567	0.587	0.516	0.458	0.542	0.484
RANK B1		7	2	1	4	6	3	5
Object management (B2)		0.357	0.645	0.326	0.565	0.366	0.635	0.635
RANK B2		6	1	7	4	5	3	2
Government services (B3)		0.769	0.649	0.506	0.430	0.566	0.390	0.611
RANK B3		1	2	5	6	4	7	3
Public safety management (B4)		0.358	1.000	0.573	0.687	0.659	0.207	0.200
RANK B4		5	1	4	2	3	6	7
Adjustments and corrections (B5)		0.605	0.449	0.748	0.252	0.662	0.551	0.662
RANK B5		4	6	1	7	3	5	2
Community culture (B6 )		0.075	0.925	0.689	0.585	0.585	0.832	0.921
RANK B6		7	1	4	5	6	3	2
Total score		2.480	4.235	3.429	3.035	3.296	3.157	3.513
Total rank		7	1	3	6	4	5	2

. According to the evaluation results, N3 had the best performance in the B1 and B5 and had the worst performance in the B2. The N2 obtained the highest score for the B2. In terms of the B1 and the B3, N1 had the best performance, while N6 had the worst. In terms of the B4, N2 obtained the highest score, and N7 obtained the lowest score. N2 obtained the highest score in the B6, whereas N1 obtained the lowest score.

The total score of the smart community governance was calculated by the above six dimensions, as shown in Table 4. The results showed that the average score of seven sample smart communities is 3.306. The range of the performance score was designed between 0 and 5. Thus, considering the value range of the performance, five grades of the overall score could be classified, namely the highest performance (4~5), high performance (3~4), general performance (2~3), low performance (1~2), and the lowest performance (0~1). The rate of smart community governance in each category could then be counted: 17% with the highest performance, 66% with a high performance, and 17% with a general performance. Although only N2 reached the highest performance, more than half of the smart communities showed a high performance. Therefore, the overall performance of smart community governance was at a relatively high level.

5. Discussion

5.1. Distributions of Evaluation Indicators Weight of SMART Community Governance

Statistically significant differences were seen in the weight of the evaluation indicator for smart community governance. On the one hand, the B5 dimension (Adjustments and Corrections) obtained the highest average weight followed by the B3 dimension (Government Services). Considerable attention has been given to the subjective well-being and life satisfaction of residents, especially the quality of community services [74,75]. The ”Adjustments and Corrections” and ”Government Services” are the important components of the smart community service. Therefore, a higher weight was assigned to the B5 and B3 dimensions. On the other hand, the average weight of the B1 (Public Security Controls) and B4 (Multiple Participation) dimensions is lower than other dimensions which may be affected by the indicator C14 and C43. The absence of social enterprises participating in the governance of community affairs often leads to these enterprises [76,77]. Meanwhile, the relationship between the police and residents is sometimes plagued by the shallow level of public relations which affected the quality of police–citizens interaction [64]. Due to these, the evaluation indicator ”The participation of social organizations governance” (C14) and “Police Citizen Interaction” (C43) received lower weights.

5.2. Differences in the Overall Level of Smart Community Governance

The results showed that the overall performance of the smart community governance was at a relatively high level, with most of the communities achieving high performance. In detail, N2 obtained the best performance overall. According to Table 4, the N2 ranked best in the B1, B2, B3, B4, and B6 dimensions, whereas it ranked worst in the B5 dimensions. This was mainly due to the low scores in the indicator ”Legal Services” (C53) which had not yet met the expectations of community governance development. Despite the online platform for legal services that had been developed in the N2, residents could not obtain online legal services through it directly. In addition, the N1 had the lowest score overall, but received the best performance in the B3 dimensions. This result was affected by the excellent performance in the ”Service for the Floating Population” (C36). There is a large floating population in Guangzhou city where the N1 is located. Through the use of rental housing registration, the ”Ping-an Baiyun” service mini-program and access control systems, the precise management of the floating population was enhanced in the N1. The difference in the governance level of smart communities indicated an unbalanced development of these smart communities and that each smart community needs a customized development plan for enhancing its governance performance.

Findings from the results can be provided as references for policymakers. For example, the N7 compared to the other smart communities was in a better situation in terms of the B2, the B3, the B5, and the B6 dimensions, but in terms of the B5 and the B6 dimensions, it performed less well. The findings provide decision-makers with a direction for improving governance performance. Decision-makers should pay more attention to the aspects which performed less well, and could conduct detailed investigations on the top-level design and platform development of the community.

5.3. Priority of Development Strategies for Smart Communities

In order to provide a priority of strategies for each smart community, further comparisons were conducted with the scores of the smart communities. The performance of smart community service and smart community management were the main components of the smart community governance [16]. Thus, we divided the performance into two parts: the performance of smart community management and the performance of smart community service.

The performance of smart community management came from the B1, the B2, and the B4, while the performance of smart community service was calculated by the B3, the B5, and the B6. Therefore, a decision-making matrix could be drawn (Figure 5). As demonstrated in Figure 5, the N2 in quadrant I showed excellent performance for both smart community management and smart community service. Further improvement strategies for these smart communities are recommended through referring to their own specific problems with considerations of other factors.

The N3, N5, N6, and N7 communities in quadrant Ⅱ performed better in smart community service, but relatively worse in smart community management than the smart community in quadrant I. This situation may be due to two reasons. On the one hand, there was insufficient participation of governance stakeholders [16,78]. The findings implied that these smart communities have not demonstrated all types of communication channels to serve the smart community governance. For example, official online participation channels for the social enterprises in governance were not provided in the N3 and N5, and similar participation channels for community organizations were also unavailable in the N6 and N7. Therefore, the platforms for communication could be encouraged and utilized across these communities to promote community engagement. On the other hand, digital management applications have not been widely accepted within smart communities, which forces smart community management to face challenges [78]. In some cases, digital management tools were insufficient in these smart communities, such as the inadequate provision of mobile office equipment for administrators. Consequently, strategies should focus on upgrading management systems and strengthening the digital infrastructure.

The N4 community in quadrant Ⅳ was better in smart community management but relatively worse in smart community service than the smart community in quadrant I. A possible explanation was the issues of mismatching services and demands in the smart community [9]. With changes in people’s lifestyles, it becomes difficult for the smart community to meet the growing service demands of residents [79,80,81]. For example, the service platform of the N4 was able to provide basic public services for older people but failed to provide online sports and recreational activities. Hence, priority should be given to the needs of residents in the construction of smart communities, and the content of smart community construction should be dynamically adjusted according to their expectations of community services [9].

The N1 community in quadrant Ⅲ was worse both in smart community management and smart community service than the communities in the other quadrants. Therefore, it is important to promote these communities in both smart community management and smart community service (the strategy of improvement can be referred to the smart communities in quadrant Ⅱ and quadrant Ⅳ).

5.4. Suggestions for Improving the Governance Performance of Smart Communities

According to the above findings, several suggestions could be obtained as references for decision-makers to improve the level of smart community governance.

First, devoting significant resources to develop smart community infrastructure could improve the level of smart community governance. A promising strategy is to provide ample smart infrastructures in smart communities by making full use of telecommunications infrastructure and community e-commerce [82]. Besides, various communication platforms could be offered to support information exchange between community stakeholders, such as establishing community WeChat groups or official community accounts to improve information communication [82,83].

Second, facilitating communication among stakeholders may lead to better performance of smart community governance. To achieve effective community consultation and participation, it is recommended that various information communication channels be established for conveying messages about smart community governance and providing diversified forms of ways to participate in the life-cycle construction of smart communities [84]. Specifically, a community information feedback channel could be optimized to help all participants follow up with smart community governance progress. Meanwhile, importance should also be attached to the maintenance of communication platforms in these communities. For instance, data from the monitoring system should be transmitted and stored in the cloud platform database to keep remote access convenient.

Third, increasing funding is essential for smart community construction. The participation of social capital in smart community governance could be encouraged to reduce financial pressure on the government. To ensure capital guarantees for smart community development, modern financial models are recommended for private enterprises to invest in the implementation of smart community governance, such as public–private partnerships [85].

In addition, the actual situation of smart community varies from place to place and the focus for smart community governance in different regions should be variable according to local conditions. Thus, it is particularly important to understand the direction of community governance work of local smart communities.

6. Conclusions

To help decision-makers make community governance development decisions for smart communities, an integrated evaluation model of smart community governance was developed and the governance levels of seven smart communities were analyzed as cases by applying this method. Several findings are worth mentioning. First, a comprehensive evaluation indicator system was established, and a systematic framework for the performance evaluation of the governance level of multiple smart communities was developed through the CRITIC-TOPSIS method. Second, the results of weights showed that the indicator of the highest weight was the ”Mediation of Conflict”, while the lowest was the ”Housing Management”. Third, the evaluation results indicated that the Yucun community obtained the best performance while the Moxi community obtained the worst performance. Furthermore, those lower ranking dimensions should be given more attention such as the ”Adjustments and Corrections” dimension of the Yucun community, the ”Adjustments and Corrections” and the ”Community Culture” dimensions of the Jinbangyuan community. This research contributes both to the innovation of community governance evaluation and to the improvement of smart communities. In practice, the governance performance of various smart communities could be assessed by the evaluation model developed in this study, and the governance levels of these communities could be promoted on this basis.

Future research could be carried out when large-scale data sets are accessible to evaluate the performance of smart community governance. Furthermore, the performance of smart community governance may change over time, and frequent evaluations are recommended for future comparisons in order to encourage continuous improvement.