Evaluation and Optimization Paths of Design Elements of Underground Building Atria Based on IPA–Kano Model

Abstract

The interior of an underground atrium is often a combination of multiple design elements. The lack of methodology and quantitative data support makes it impossible for developers and operators to know the public’s preferences for further renovation, which seriously hinders the sustainable development of underground atria. This paper addresses this issue by evaluating the IPA–Kano model for underground atria with three cases selected in the main urban area of Chongqing for comparison. Then, three types of design elements have been identified to have different impacts on underground building atria on the basis of analyzing the relationship between design elements and users’ subjective perceptions of underground building atria. They are the basic, importance, and attraction elements. Finally, the optimization path for underground building atria is determined by integrating the actual performance of these three types of design elements. It is indicated that design elements such as seating, interface, and signage must be prioritized for improvement.

1. Introduction

POPS (Privately owned public spaces) was first introduced in New York in the 1960s and is defined as a hybrid space owned and managed by the private sector but open to the public for free public use [1,2]. POPS has three functions: firstly, it relieves strained municipal resources for local authorities [1,3,4,5,6]. Secondly, POPS provides more potential consumers and additional economic benefits for private developers [2]. Lastly, and most importantly, systemic and high-quality POPS design can provide a mix of spatial features and comfortable spatial experience, directly benefiting the users [7,8].

As an essential node space of POPS, the atrium of an underground building plays a vital role in creating an excellent psychological environment and enhancing the overall image of the space due to its public and decentralized use, the uniqueness of its spatial composition, and its closeness to natural light and natural landscape for underground buildings with few floors and a relatively large plan volume. However, it can be found from previous studies that discussions mainly revolve around clarifying the significance of an underground building atrium within the whole underground building. For example, from the perspective of design, with the underground building atrium as an underground building orientation, Jian Zhang et al. proposed a spatial image design countermeasure using nodes such as the underground building atrium for the soft orientation system of the underground complex [9]. Through field research, Jia Mi et al. found that the underground building atrium, as an important node for the connection between the underground building and the above-ground environment, could help users to find their way and identify their direction in the underground building [10]. From the perspective of interactive penetration between above ground and underground, Jian Zhao et al. argued that the arrangement of underground building atria could not only strengthen the physical connection of the internal space, but also play a vital role in promoting the visual penetration of the space [11]. Taek et al. argued that the atrium of an underground building, as an intermediary space connecting above and below ground, could guide customers to spend more time in the underground shopping center [12]. From the angle of fire evacuation function, Jian Zeng and Qiao Wang et al. illustrated the role of fire prevention and evacuation of an underground building atrium and argued that it should be used as a buffer zone for horizontal and vertical evacuation [13]. To sum up, it can be seen that the underground atrium plays a crucial part in the whole POPS, but there are few quantitative studies on how to optimize the underground atrium at this stage.

As an essential underground public space, the largest user group of underground building atria is the public, and the solution to user needs should be prioritized for optimization. At present, as a popular technology in evaluating customer needs, the IPA–Kano model is used to assess, by identifying the relationship between service quality and user needs, and hence propose an optimization path according to the results. However, the interdisciplinary application of the model in urban research is limited to traffic services, pedestrian satisfaction, noise, and other fields [14,15]. However, it is seldom applied in POPS studies. This study plans to extract the implicit importance of the design elements through bivariate correlation analysis by describing and analyzing the attribute performance and significance of the design elements of underground building atria in the main urban area of Chongqing. Then, the design elements of underground building atria are evaluated using the IPA–Kano model, and the optimization path is analyzed based on the measurement results combined with the actual performance of the attributes.

2. Method

2.1. IPA–Kano Model

For the choice of spatial element quantification methods, the satisfaction measurement tools of IPA (Importance-Performance Analysis) [16,17] and the Kano hierarchy of needs model [18,19] are most commonly used. IPA was developed by Martilla and James to provide insight into the elements affecting users’ evaluation of overall performance and to judge their priorities [20]. The means of satisfaction and perceived importance of each IPA indicator are projected in the form of coordinates onto four quadrant diagrams, and elements falling into different quadrants have different connotations (Figure 1). However, in practice, the relationship between element attribute performance and overall performance is nonlinear and asymmetric [21]. This problem can be better addressed using the Kano model proposed by Japanese scholar Noriaki Kano, which classifies user needs into five levels: attractive needs, desired needs, necessary needs, undifferentiated needs, and reverse needs. Elements under different need level categories have different degrees of influence on the overall performance (Figure 2), but the model itself cannot directly derive policy recommendations. In addition, undifferentiated and reverse demands do not contribute much to the overall performance so their characteristics are not involved in our study. In summary, a combination of both needs is to be applied in the evaluation.

Kurt Matzler et al. proposed the IPA–Kano model by integrating them, and it has become a popular evaluation technique in urban management and product services, and is also gradually being introduced into urban planning and architectural design [20,22]. It combines the advantages of IPA in judging the priority of relevant elements with the merits of Kano in evaluating the attributes of user needs, with the horizontal coordinate representing the explicit importance and the vertical coordinate representing the implicit significance. The importance elements in quadrant A have the highest explicit and implicit importance. Quadrant B is the attractive element with low explicit importance and high implicit importance. Quadrant C is the unimportant element with the lowest explicit and implicit importance. Quadrant D is the basic element with high explicit importance and low implicit importance (Figure 3). The model provides a prediction of user needs and causal path analysis, which offers guidance for measuring design elements and proposing improvement strategies.

2.2. Optimization Method

Our research focuses on the basic elements, importance elements, and attraction elements. The most significant is the linear effect of importance elements on the overall performance that can be improved if they are delivered and vice versa. Basic elements do not affect the overall performance if they are satisfied but will degrade it if they are not. The delivery of attraction elements will improve the overall performance, otherwise it will not be contaminated. Whether the remaining non-important performance elements are satisfied or not will not affect the overall performance so they are not of concern [15].

Although the evaluation results of the design elements of underground building atria can be derived, these are insufficient to provide an optimization path for their development. Since the contribution of design elements to the overall performance depends on their actual performance, the evaluation results should be combined with the attribute performance to determine the optimization path for the underground atrium design elements. It is important to note that the effects of basic and attraction elements on the overall performance are asymmetric. For instance, the overall performance is not enhanced with basic elements exceeding expectations, but gets degraded if not implemented. Attraction elements do not cause a decrease in the overall performance if they are not realized, but can improve it if realized. The impact of important performance elements on the overall performance is proportional to the degree to which they are realized. Thus, the order is basic elements, importance elements, and attraction elements [23]. The actual performance of the design element attributes is divided into three categories (

Table 1. Improvement priority.

Improvement Priority	Quadrant	Attribute Performance Ranking
1	Basic Element	22–33
2	Important element	22–33
3	Attractive element	22–33
4	Basic Element	11–22
5	Important element	11–22
6	Attractive element	11–22

), reflecting different performance levels. The first 11 are the best performers and the last 11 are the worst, while the others are relatively mediocre. We can derive the optimization paths by constructing six priority levels.

2.3. Experimental Process

The experimental process consists of the following four steps (Figure 4).

(1)

We conducted and collated questionnaires to calculate the mean importance and attribute performance values of atrium design elements of underground buildings in the three cases.

(2)

The importance and attribute performance bivariate was correlated using spss22.0 data statistical analysis software to extract the implied importance of atrium design elements of underground buildings [24].

(3)

The IPA–Kano analysis raster was used to present the comparative results of explicit and implicit importance, using importance data to represent explicit importance on the horizontal axis and bivariate correlation analysis data of importance and attribute performance to describe implicit importance on the vertical axis. The mean value of the 33 elements of explicit and implicit importance was used as the central coordinate, and each element corresponded to the quadrant diagram, respectively.

(4)

Based on the results of the quadrant attribution of the atrium design elements of underground buildings, the optimization path was analyzed by combining the performance ranking of each element attribute.

Among them, SP22.0 was adopted to perform bivariate correlation analysis to extract the implied importance of the atrium design elements of underground buildings with the following calculation formula.

$$r_{xy}=\frac{\sum x_i{y}_i-n\overline{xy}}{(n-1)s_x{s}_y}=\frac{n\sum x_i{y}_i-\sum x_i\sum y_i}{\sqrt{n\sum x_i^2-{\left(\sum x_i\right)}^2}\sqrt{n\sum y_i^2-{\left(\sum y_i\right)}^2}}$$

The coefficients take values between −1.0 and 1.0, with variables close to 0 being no correlation and close to 1 or −1 being a strong correlation.

3. Indicator and Data

3.1. Design Elements of Underground Building Atria

An atrium usually refers to the courtyard space inside a building; a unique form that is both isolated from and integrated with the external space. In underground buildings, the application of an atrium can solve such problems as undesirable psychological reactions, inconspicuous external images and features, limitations of view and natural light, and poor sense of direction. Therefore, “the atrium of an underground building (hereafter referred to as atrium) is a diverse, evolving and inspiring dynamic space that contains humans, activities and natural elements” [25]. The atrium is characterized by the following points. (1) First of all, it can often become a transportation hub or “central landmark” for the whole underground building, and its multifunctional internal compound provides a platform for public socialization and recreation. (2) Unlike general underground spaces that are in a completely enclosed environment, its top often takes the form of an open hole as a way to provide natural light. Meanwhile, when the atrium is composed of multiple layers, the inner facade often uses windows and doors and corridors to form a staggered facade level. (3) Because of the good natural environment, greening and resting facilities are often set here. It can be seen that the atrium has characteristics different from other types of underground public spaces, which are reflected in the design elements of public facilities, physical environment, spatial interface, and spatial form. We conducted an in-depth interview with the head of the property department of Nanping Yixiang City in Chongqing, and concluded that from the perspective of operation and management, public facilities should be the most important, followed by spatial interface and physical environment, and spatial scale and morphology are relatively less important. This is also a reasonable speculation, because the public can first perceive public facilities [26], while other design elements are relatively weakly perceived.

Then, based on the perspective of architecture, this study combines the features of atria as an essential node of above-ground penetration and a key shared underground space [27,28] with its basic functional units such as gathering, dispersal, and public activities, as well as the constituent elements of morphology, interface, environment, and facilities. The four indicators are proposed: B1 (spatial scale and morphological elements), B2 (spatial limited degree elements), B3 (spatial physical environment elements), and B4 (spatial facilities elements). Further subdivided under the second level, 33 design elements of an atrium are finally summarized, covering many aspects such as guidance modes, spatial combination forms, environment, and service facility layout (

Table 2. Design elements of the atrium space of the underground building.

B1 Spatial scale and form elements	C1 Spatial scale	D1 Plane ratio
		D2 Section ratio
	C2 Spatial form	D3 Plane layout
		D4 Atrium Sectional Form
B2 Spatial limit elements	C3 Top interface	D5 Roof opening diameter
		D6 Roof shape
		D7 Roof transparency
	C4 Vertical interface	D8 Wall form change
		D9 Ratio of door and window openings
		D10 Fence and railing form
	C5 Bottom interface	D11 Degree of ground undulation
		D12 Balcony form
B3 Spatial physical environment elements	C6 Light environment	D13 Artificial lighting
		D14 Natural lighting
	C7 Sound environment	D15 Sound comfort
		D16 Sound preference
	C8 Ventilation	D17 Ventilation
	C9 Thermal environment	D18 Temperature and humidity
B4 Space facility elements	C10 Landscape facility	D19 Wall material and color
		D20 Floor pavement pattern
		D21 Plane greening
		D22 Vertical greening
		D23 Landscape furniture
		D24 Water feature
	C11Transportation facility	D25 Escalator setup
		D26 Elevator setup
		D27 Stair Step Dimensions
		D28 Ramp shape and width
		D29 Winding corridor, verandah form
	C12 Guide facility	D30 Map marking
		D31 Underground traffic sign
	C13 Rest facility	D32 Form of rest seat
		D33 Number of rest seat

). Importance and attribute performance questionnaires are developed based on these 33 design elements.

3.2. Questionnaire Design

The importance questionnaire is a 5-point Likert scale with 33 questions based on user preferences for atrium design elements. For instance, in D33 it asks “Do you think the number of lounge seats is important?” where “1” means very unimportant, “2” unimportant, “3” average, “4” important, and “5” very important. The weighted average score of each variable is adopted as the importance score to measure the explicit importance of each design element of atria. The attribute performance questionnaire is designed for users to rate the actual performance of the design elements of atria on a 5-point Likert scale. For example, in D33 it asks “Are you satisfied with the current number of lounge seats?” where “1” means very dissatisfied, “2” dissatisfied, “3” average, “4“ satisfied, and “5” very satisfied. The weighted average score of each variable is adopted as the attribute performance score, reflecting the actual performance of the design elements of atria. The importance and attribute performance questionnaires not only ask questions about the 33 elements but also record the age and gender of the participants.

3.3. Research Cases and Data

Chongqing is one of the four municipalities directly under the central government in China, and the average density of the main urban area is about 20,000 people/km² [13]. With an early utilization of urban underground public space, it is abundant in type and quantity. We selected three atria as evaluation cases, respectively, in Nanping Yixiangcheng (hereinafter referred to as Nanping), Jiefangbei PARK108 Guotai Youhe City Plaza (Jiefangbei), and Guanyinqiao Jinyuan Buoyant City (Guanyinqiao). These three cases are located in the main urban area of Chongqing, all within the business districts, and the internal underground atria are located at the core of the whole space, as well as having a high pedestrian flow (Figure 5).

The study was conducted from 2019 to 2021 using questionnaire distribution. Offline was primary and online secondary. Offline questionnaires were mainly distributed in the underground atria. An online questionnaire survey was conducted using the “Sojump” platform, and distributed to former users or users through social networks, e-mail, and other means. A total of 689 questionnaires were returned and sorted. On the one hand, the questionnaires with obvious and illogical problems were eliminated. On the other hand, a quota sampling was used for the age structure of the questionnaire to ensure age distribution, with each of the four age groups accounting for 1/4 of the total number of participants (

Table 3. Participant age and gender ratio.

Age				Gender
18–25	26–35	36–45	Over 46	Male	Female
25%	25%	25%	25%	42%	58%

). Finally, 659 valid questionnaires were collected, 222 from Nanping, 222 from Jiefangbei, and 215 from Guanyinqiao. Notably, 58% were from females and 42% were from males, indicating a reasonable distribution because underground public spaces are mainly commercial and tend to be more popular with women [29]. Then, the reliability and validity of the 33 elements in the questionnaires were analyzed using spss 22.0 statistical software, showing the Cronbach values were all above 0.8, which proved that the questionnaires had good reliability. The significant values were all 0, meaning good validity (

Table 4. Reliability and validity analysis of the questionnaire.

Questionnaire	Name	Nanping	Jiefangbei	Guanyin Bridge
Attribute Performance	Cronbach Alpha value	0.925	0.94	0.94
	Number	33	33	33
Importance	Cronbach Alpha value	0.963	0.878	0.884
	Item number	33	33	33
Attribute Performance	Kaiser-Meyer-Olkin	0.882	0.901	0.898
	Bartlett Sphere Test	0	0	0
Importance	Kaiser-Meyer-Olkin	0.925	0.754	0.825
	Bartlett Sphere Test	0	0	0

).

4. Results

4.1. Importance and Attribute Performance Analysis

In terms of the importance of the design elements of the atria, different age groups all consider ventilation, temperature, and humidity to be very important. However, the female group believes that in addition to the two elements mentioned above, the light environment, especially artificial lighting, is of high importance, while males believe that this element is of general importance. In general, it is agreed by users that ventilation, temperature, and humidity are very significant design elements in the physical environment of the space, while changes in the form of walls and eaves are considered unimportant (Figure 6). The importance of other identical elements varies significantly among users. ANOVA tests on the same design elements in the three cases yield remarkable p-value levels, 11 of which are significant (Figure 7). Plane greening, water features, stair step dimensions, and ramp shape and width are more crucial in Guanyinqiao than in Jiefangbei and Nanping, while vertical greening is less important than these elements in the latter two. The fence and railing form, elevator setup, and winding corridor and verandah form in Guanyinqiao are more important than those in Jiefangbei.

In terms of the attribute performance of the atrium design elements, there are significant differences in the ratings of different age groups and gender groups. For example, the number of rest seats: those 46 years old and above and the male group feel unsatisfied, while other age groups and females are generally satisfied. In general, what users perceive to be better are ventilation, temperature, humidity, wall material and color, floor pavement pattern, and landscape facility, while ramp shape and width are poor (Figure 6). ANOVA tests on the same elements in the three cases present p-values at significant levels, with 24 items showing significance (Figure 7). All design elements in Guanyinqiao have displayed better performance than those in the other two, except vertical greening and seat number. The design elements significantly better in Nanping than in Jiefangbei are the degree of ground undulation, plane greening, landscape furniture, water features, stair step dimensions, map marking, and rest seating. The only element significantly poorer than that in Jiefangbei is vertical greening, while the performance of other elements is not very different.

4.2. Extraction of Implicit Importance of Design Elements

Explicit importance refers to users’ direct evaluation of the importance of the design element, while implicit importance refers to the importance status reflected in users’ evaluation of other aspects (

Table 5. Implicit importance and ranking of atrium design elements in underground buildings.

	Guanyin Bridge		Jiefangbei		Nanping
	Score	Rank	Score	Rank	Score	Rank
D1 Plane ratio	0.688	2	0.567	9	0.580	19
D2 Profile ratio	0.594	8	0.466	19	0.606	16
D3 Plane layout	0.517	14	0.542	11	0.721	1
D4 Atrium Sectional Form	0.661	4	0.425	22	0.636	10
D5 Roof opening diameter	0.601	7	0.489	17	0.584	17
D6 Roof shape	0.683	3	0.718	1	0.681	4
D7 Roof transparency	0.628	5	0.692	2	0.634	12
D8 Wall form change	0.612	6	0.542	12	0.673	5
D9 Ratio of door and window openings	0.521	13	0.494	16	0.622	15
D10 Fence, railing form	0.532	12	0.558	10	0.383	28
D11Degree of ground undulation	0.561	10	0.225	32	0.721	2
D12 Balcony form	0.358	28	0.357	25	0.366	29
D13 Artificial lighting	0.492	20	0.439	20	0.577	20
D14 Natural lighting	0.517	15	0.601	5	0.535	24
D15 Sound comfort	0.696	1	0.617	3	0.635	11
D16 Sound preference	0.515	16	0.596	6	0.644	8
D17 Ventilation	0.513	17	0.572	8	0.665	6
D18 Temperature and humidity	0.541	11	0.603	4	0.626	14
D19 Wall material and color	0.593	8	0.531	14	0.583	18
D20 Floor Pavement Pattern	0.469	22	0.427	21	0.686	3
D21 Plane greening	0.486	21	0.377	24	0.561	22
D22 Vertical greening	0.199	32	0.290	29	0.654	7
D23 Landscape furniture	0.467	23	0.207	33	0.576	21
D24 Water feature	0.374	26	0.344	26	0.192	33
D25 Escalator setup	0.398	25	0.323	28	0.308	31
D26 Elevator setup	0.341	30	0.534	13	0.336	30
D27 Stair Step Dimensions	0.513	18	0.330	27	0.630	13
D28 Ramp shape and width	0.258	31	0.379	23	0.384	27
D29 Winding corridor, verandah form	0.347	29	0.588	7	0.392	26
D30 Map Marking	0.452	24	0.489	18	0.269	32
D31 Underground traffic sign	0.508	19	0.510	15	0.642	9
D32 Form of rest seat	0.361	27	0.244	31	0.560	23
D33 Number of rest seat	0.194	33	0.253	30	0.439	25

). Explicit and implicit are interrelated but not identical so they must be compared and combined when evaluating design elements. The p-value for the same design element is less than 0.05, which is statistically significant, and the results and score rankings are shown in Table 4. In the three cases, the implicit importance rankings for the roof shape, ratio of door and window openings, balcony form, and artificial lighting are basically the same, while the implicit importance rankings for other elements are significantly different. Sound preference and ventilation rank highly in Jiefangbei and Nanping, while ranking moderately in Guanyinqiao.

4.3. Evaluation of Design Elements of Underground Building Atria

The same design elements in different construction situations can have diverse impacts on users’ subjective perceptions. Conversely, different design elements may have the same effect on users’ subjective perceptions in the same construction situation [14]. In the three cases, similarities and differences coexist in the evaluation results of the design elements of atria (Figure 8,

Table 6. Evaluation results of design elements of underground building atria.

Attribution of Elements	Guanyin Bridge	Jiefangbei	Nanping
Basic element	D30 Map Marking	D33 Number of rest seat	D30 Map Marking
	D33 Number of rest seat	D11Degree of ground undulation	D14 Natural lighting
	D25 Escalator setup	D13 Artificial lighting	D31 Underground traffic sign
	D28 Ramp shape and width	D20 Floor Pavement Pattern
	D22 Vertical greening	D32 Form of rest seat
Important element	D3 Plane layout	D3 Plane layout	D3 Plane layout
	D7 Roof transparency	D7 Roof transparency	D7 Roof transparency
	D16 Sound preference	D16 Sound preference	D16 Sound preference
	D17 Ventilation	D17 Ventilation	D17 Ventilation
	D18 Temperature and humidity	D18 Temperature and humidity	D18 Temperature and humidity
	D19 Wall material and color	D19 Wall material and color	D19 Wall material and color
	D31 Underground traffic sign	D31 Underground traffic sign	D31 Underground traffic sign
	D10 Fence, railing form	D1 Plane ratio	D1 Plane ratio
	D13 Artificial lighting	D2 Section ratio	D2 Section ratio
		D15 Sound comfort	D15 Sound comfort
		D9 Ratio of door and window openings	D9 Ratio of door and window openings
		D10 Fence, railing form	D13 Artificial lighting
		D5 Roof opening diameter	D4 Atrium Sectional Form
		D14 Natural lighting	D6 Roof shape
		D26 Elevator setup	D11 Degree of ground undulation
		D29 Winding corridor, verandah form	D22 Vertical greening
		D30 Map Marking	D20 Floor Pavement Pattern
			D27 Stair Step Dimensions
			D32 Form of rest seat
			D33 Number of rest seat
Unimportant element	D12 Balcony form	D12 Balcony form	D12 Balcony form
	D24 Water feature	D24 Water feature	D24 Water feature
	D21 Plane greening	D21 Plane greening	D25 Escalator setup
	D23 Landscape furniture	D23 Landscape furniture	D26 Elevator setup
	D26 Elevator setup	D25 Escalator setup	D29 Winding corridor, verandah form
	D29 Winding corridor, verandah form	D4 Atrium Sectional Form	D10 Fence, railing form
	D20 Floor Pavement Pattern	D22 Vertical greening	D28 Ramp shape and width
	D32 Form of rest seat	D27 Stair Step Dimensions
		D28 Ramp shape and width
Attractive element	D8 Wall form change	D8 Wall form change	D8 Wall form change
	D6 Roof shape	D6 Roof shape	D5 Roof opening diameter
	D1 Plane ratio		D23 Landscape furniture
	D2 Section ratio		D21 Plane greening
	D5 Roof opening diameter
	D4 Atrium Sectional Form
	D11 Degree of ground undulation
	D9 Ratio of door and window openings
	D14 Natural lighting
	D15 Sound comfort
	D27 Stair Step Dimensions

). The similarities are reflected in the fact that the same design elements belong to the quadrant of the same category. The number of rest seats is common to Guanyinqiao and Jiefangbei, while map marking is common to Guanyinqiao and Nanping. Regarding important expressive elements, Guanyinqiao has a much lower number of elements than Jiefangbei and Nanping. Plane layout, roof transparency, sound preference, ventilation, temperature and humidity, wall material and color, and underground traffic signs are important performance elements common to all three cases. Jiefangbei and Nanping have the same plane ratio, profile ratio, and sound comfort ratio to the wall opening. Guanyinqiao and Jiefangbei share the fence and railing form, and Guanyinqiao and Nanping share the artificial lighting elements. In terms of attraction elements, the only element common to all three cases is the wall form change, while the element common to Guanyinqiao and Jiefangbei is the roof shape. The other elements of Guanyinqiao are significantly more than those of Nanping and Jiefangbei.

The differences are reflected in the fact that the same design elements are grouped into different quadrants (Figure 8, Table 6). For instance, among the basic elements, Jiefangbei has a degree of ground undulation, artificial lighting, floor pavement pattern, and forms of rest seating, which are classified as important elements in Nanping. Meanwhile, natural lighting and underground traffic signs are classified as important elements in Jiefangbei. The basic elements such as elevator setup and winding corridor and verandah form of Guanyinqiao are classified as non-important elements in Nanping and Jiefangbei, while the exclusive basic elements in the latter two, except underground traffic signs and artificial lighting, are classified as non-important and attraction elements in Guanyinqiao. With regard to the importance elements, the elevator setup and the winding corridor and verandah form of Jiefangbei are classified as non-important elements in Nanping, while the roof opening diameter is classified as an attraction element in Nanping. The vertical greening, atrium sectional form, and stair step dimensions in Nanping are classified as unimportant expressive elements in Jiefangbei, while the latter two are classified as attraction elements in Guanyinqiao. Guanyinqiao presents more attraction elements, while plane ratio, profile ratio, ratio of door and window opening, and sound comfort, as mentioned above, are classified as important elements in Nanping and Jiefangbei. Nanping’s exclusive landscape furniture and plane greening are classified as unimportant elements in Jiefangbei and Guanyinqiao. This suggests that there is an apparent discrepancy in user preferences, which leads to a significant difference in the evaluation results of the same design elements for different atria.

5. Discussion

There are homogeneous evaluation results among different atria so we should pay attention to all these common elements (

Table 7. Improvement priorities of atrium design elements of underground buildings.

	Guanyin Bridge	Jiefangbei	Nanping
Priority 1	D33 Number of rest seat	D11Degree of ground undulation
	D25 Escalator setup	D33 Number of rest seat
	D28 Ramp shape and width
	D22 Vertical greening
Priority 2	D31 Underground traffic sign	D30 Map Marking	D11 Degree of ground undulation
		D10 Fence, railing form
		D26 Elevator setup
		D29 Winding corridor, verandah form
Priority 3	D8 Wall form change		D8 Wall form change
	D5 Roof opening diameter		D21 Plane greening
Priority 4	D30 Map Marking	D32 Form of rest seat	D30 Map Marking
			D14 Natural lighting
			D31 Underground traffic sign
Priority 5	D3 Plane layout	D3 Plane layout	D3 Plane layout
	D13 Artificial lighting	D14 Natural lighting	D15 Sound comfort
		D15 Sound comfort	D31 Underground traffic sign
		D31 Underground traffic sign	D33 Number of rest seat
		D2 Section ratio	D6 Roof shape
		D5 Roof opening diameter	D1 Plane ratio
		D9 Ratio of door and window openings	D4 Atrium Sectional Form
			D27 Stair Step Dimensions
			D9 Ratio of door and window openings
Priority 6	D11 Degree of ground undulation	D8 Wall form change	D31 Underground traffic sign
	D14 Natural lighting	D6 Roof shape
	D15 Sound comfort
	D6 Roof shape
	D1 Plane ratio
	D2 Section ratio
	D4 Atrium Sectional Form
	D27 Stair Step Dimensions

). The design elements related to spatial scale and limits affect users’ spatial experience, but are far less important than the physical environment and amenities, which are more advanced needs to be considered after the basic needs are met. The number of rest seats, degree of ground undulation, wall form change, map marking, and underground traffic signs are ideal for this purpose. The reasons for optimization are as follows: owing to the public space property of the atria [26], users want to rest and communicate here, but the number of seats is often insufficient to meet the needs of the crowd. The wall form change and the degree of ground undulation are the basis of the spatial limitation of atria, which are the design elements visible and directly perceived by the users. Unlike above-ground atria, the spatial interface in the atria often lacks variation at this stage. In addition, basic guidance facilities such as map marking and underground traffic signs also need attention. Although an atrium can get external views and better lighting through an opening in its roof, it is undeniable that there is still a lack of spatial direction reference, coupled with the fact that it is often the gathering and dispersal place of various underground channels. Therefore, a clear guiding sign plays a vital role. An atrium itself is a particularly good physical space environment, but it can be seen that the actual performance of natural lighting and sound comfort still cannot meet the expectations of users. The plane layout is an important performance-oriented element in the three places, but the actual performance is average. The roof shape is an attraction element in Guanyinqiao and Jiefangbei, reflecting a high-level demand that needs considering only after the basic demand is satisfied, hence in line with the demand hierarchy theory.

In addition to the design elements common to all three cases, the design elements to be optimized in each case are different (Table 7). The dissatisfaction of users in Guanyinqiao focuses on the design of transportation infrastructure and spatial scale. In fact, the shortcoming in the design of transportation infrastructure in an atria has been a common problem, such as the location and arrangement of escalators, the width of ramps, and the design of stair treads. Other important performance elements such as artificial lighting and top openings are of average performance in practice. It is also found in the site study that the slightly smaller diameter of the top opening and the lack of artificial auxiliary lighting have resulted in a darker overall space, and that the lower height inside Guanyinqiao exacerbates the user’s sense of oppression (Figure 9). As a unique design element of Guanyinqiao, vertical greening is also noteworthy. In Jiefangbei, it can be seen that all the other design elements can be grouped into those of a façade form, except the seat shape as a fascinating element, so the lack of variation in the spatial interface is the main reason for user dissatisfaction (Figure 10). Nanping also has the same problem as Guanyinqiao in terms of spatial scale, mainly focusing on the narrow opening and the lack of variation in the form of contours. Nanping introduces the above-ground landscape into the underground, but the attraction elements of ground-level landscape and vignette design are not set up, resulting in the lack of good viewing effect on the horizontal sight line (Figure 11).

6. Conclusions

Three cases in the main urban area of Chongqing are taken as the research object, and the IPA–Kano model is adopted to identify three types of elements that have different impacts on the atria, namely, the basic elements with high concern, the important elements that can be strengthened during the optimal allocation of resources, and the attraction elements that can be intensified after meeting the above two requirements. By integrating these three types of elements and their actual performance, we have established the optimization path for the design elements of atria. Design elements such as the number of rest seats, wall form change, and degree of ground undulation are priority elements to ensure enhanced user satisfaction. Guanyinqiao should attach importance to the transportation infrastructure issues and improve the sense of oppression caused by inadequate spatial scale and lighting through design means. Jiefangbei needs to be improved in terms of spatial interface variations. Nanping should also pay attention to the spatial scale, while appropriately shaping some ground and vertical landscapes to meet users’ viewing needs.

The IPA–Kano model adopted in this paper and the conclusions can help developers and designers understand their projects’ operational status and public preferences, as well as the key elements that affect the overall performance of underground buildings, providing data and scientific support for further project improvement and resource optimization. Furthermore, the information can be shared with building investors, who also have a vested interest in the sustained success of underground atria.

The research method in this paper has some limitations: similar to the traditional IPA importance analysis, the raster quadrant to which an element belongs depends on the position of the X and Y axes, and the choice of the threshold value affects the classification of the element category [30]. The closer the design element is to the axis, the worse its attribution. Due to the multi-collinearity among design elements, binary correlation analysis is adopted to derive their implied importance instead of the usual partial correlation analysis and regression analysis.