Impacts of Micro-Scale Built Environment Features on Tourists’ Walking Behaviors in Historic Streets: Insights from Wudaoying Hutong, China

Abstract

The impact of built environment features on tourists’ walking behaviors has received growing attention. Although many researchers have observed the effects of micro-scale factors, the impact of culture-related factors on walking behaviors has been frequently overlooked. Therefore, it is vital to synthesize those micro-scale variables to develop a more holistic picture, and incorporating a cultural perspective is an imperative for the preservation and vitality enhancement of historic streets. In our study, a micro-scale built environment (MiBE) variable system was constructed to capture the features of historic streets, and 109 visitors were tracked in Wudaoying Hutong to record their walking-stopping behaviors. The results revealed four primary components affecting walking-stopping behaviors, among which transparency was the most influential factor, followed by the transitional space between streets and buildings, contributing to 49.8% and 21.6%, respectively. Notably, the non-negligible impact of two culture-related factors, including the contrast between Chinese and Western styles and traditional Chinese features, was also revealed, contributing to 28.6% of the total observed activities. We further compared four different types of micro-scale factors of the built environment and the corresponding walking-stopping behaviors, providing both scientific and theoretical reflections for preserving and renewing historic streets.

1. Introduction

Since the mid-1990s, walkability has received extensive attention as a healthy and sustainable transport approach [1]. For one thing, no one can deny that walking is essential to enhance urban citizens’ physical and mental health, since obesity, cardiovascular diseases, stress and other mental illnesses can be substantially reduced by moderate walking behaviors [2]. In addition, walking can boost urban vitality and strengthen social interaction amongst urban residents [3,4]. In this way, walkable places are considered to be beneficial for not only real estate but also public health [5].

The term ‘walkability’ refers to how the built environment enables pedestrians’ walking behaviors, signifying that the physical environment can significantly impact tourists’ activities, positively or negatively [6]. Numerous studies have explored how macro-scale features of the built environment can promote walking behaviors, including land use, accessibility, high densities, connectivity of streets, and others [7,8,9,10]. Rodríguez et al. found that higher levels of mixed land use correlated with more sustainable travel behaviors [11]. Vergel-Tovar and Rodriguez detected similar results in the Latin American context [12]. In these studies, walking is often regarded as a mode of transportation rather than a recreational activity.

Further, plenty of studies have paid attention to variables affecting walking behaviors for recreational purposes, and most of them can be classified as micro-scale. Street greenery, street scale, and public amenities have been demonstrated to exert a beneficial effect on the walking intention of the elderly [13] and influence the recreational walking experience of visitors [14,15,16,17,18]. Despite all these studies, many micro-scale features that might influence walking-stopping behaviors have not yet been fully explored, for example, building materials and color, as well as forms of architecture. These features can be very visually attractive to walking tourists, making them stay for aesthetic and entertaining reasons [19,20]. Further, there is a lack of a systematic framework of micro-scale attributes that might impact tourists’ walking-stopping behaviors, which hinders us from fully comprehending the internal relations of those variables [1,9,21]. Understanding the effects of micro-scale variables is crucial for urban planners, architects, and landscape designers as improving those micro-scale features may result in an increased number of sightseers, thereby contributing to the vitality and sustainability of walking spaces [18,22].

Historic streets can provide an ideal space for studies on micro-scale elements of the built environment and their impact on visitors. First, as a common type of linear space in urban areas, diversified behaviors of tourists and abundant micro-scale environmental features can be easily observed and assessed in these streets. Furthermore, historic streets can serve as a space for multiple forms of social communication and interaction, such as greetings and casual conversations, which means that various types of social behaviors can be detected [23]. Additionally, buildings, stores, decorations of different ages, and styles can often be found in historic streets, displaying diverse cultural elements of the built environment. These culture-related factors are not readily captured in other public spaces [23]. Existing studies on the impact of micro-scale features of built environments have been mainly conducted in commercial blocks. However, the investigation of cultural characteristics should never be neglected; therefore, the in-depth analysis in historic streets is urgently required.

To address these gaps in existing literature, the overarching objective of this paper is to assess the association between micro-scale characteristics of built environments and tourists’ walking behaviors in the context of historic streets. Understanding this correlation is not only crucial for perceiving tourists’ travel behaviors, but also for the planning and architectural design of business streets with historic characteristics. To achieve the overall objective, this article pursues two research questions:

(1) What micro variables of the built environment can attract the attention of tourists and trigger walking-stopping behaviors?

(2) To what extent can these micro-scale elements of built environment affect tourists’ walking-stopping behaviors?

To answer these questions, a micro-scale built environment (MiBE) variable system was constructed in this study based on a literature review, after which the correlations between micro variables and the walking behaviors of tourists were investigated by principal component analysis and multiple linear regression analysis.

The remainder of this paper is organized as follows. Section 2 reviews the micro-scale elements of the built environment and tourists’ walking behaviors in historic spaces. Section 3 illustrates the study area and explains how the microscopic built environment variables are selected, collected, and processed. Section 4 presents the descriptive statistics of these collected micro-scale features and identifies vital impacting factors, which are later discussed in Section 5 to provide some implications for architects and urban planners. The final section concludes the main findings of this study and points out its limitations as well as areas for future research.

2. Literature Review

2.1. Micro-Scale Built Environment Elements

As depicted in

Table 1. Micro-scale elements in the existing literature.

Space	Feature	Element	Case Type	Case Location	Literature
Street	Street scale	Width, space enclosure	Commercial streets around subway stations	Shanghai, China	[34]
		Integration, orderliness,	Various streets	Montreal and Toronto, Canada	[18]
	Infrastructure, sign	Traffic sign, vehicle, construction facility, object	Neighborhood streets	Bunkyo Ward, Tokyo, Japan	[17]
		signs and public service facility number	Neighborhood streets	Henan, China	[35]
		Benches, plant characteristics	Commercial streets	Guangxi, Guangdong, China	[29]
		Services, hotel and catering infrastructure, historic buildings	Rural communities	Brzeski, Poland	[25]
	Public space	Sidewalk, Ppublic space area	Commercial blocks	Da Nang, Vietnam	[30]
	Aesthetics	Space quality, spatial, historical atmosphere, landscape design	Historic streets	Beijing, China	[31]
		Aesthetic perception	Rural communities	Mississippi, USA	[27]
		Perceived colors	Commercial streets	Paris, France	[24]
Transition space	Interface Ground materials	Permeability	Commercial streets around subway stations	Shanghai, China	[16]
		Floor covering materials	Commercial streets	Zhejiang, China	[32]
Building	Ground-floor interface	Transparency	Commercial streets	Shanghai, China	[33]
		Imageability, complexity	Neighborhood streets	Salt Lake City, Utah, USA	[26]
		Coherence, legibility, mystery	Historic streets	Taiwan, China	[28]

, micro-scale elements of the built environment that influence travel behaviors can be briefly classified into three types based on their spatial attributes: street spaces, transitional spaces, and buildings (Table 1). These studies have been carried out on different types of streets and in various countries across Europe [24,25], America [18,26,27], and Asia [16,17,28,29,30,31,32,33,34,35].

The four dominant factors influencing visitors’ behaviors in street spaces in existing literature include the street scale, public infrastructure, public space, and aesthetic elements. Gehl [14], Chen [34], and Steinmetz-Wood [18] noted that length, width, integrity, and orderliness could exert different impacts on the experience of pedestrians, and these findings have been confirmed in many metropolises, such as Tianjin, Montreal, and Toronto. Furthermore, public infrastructure of multiple types and forms, such as benches, lighting, trees, and traffic signs, are appealing to travelers [17,29,35]. For example, the impact of trees on visitors’ physical comfort and psychological well-being is widely recognized [36]. Thirdly, the quality of public spaces, such as the sidewalk flatness and neatness, is repeatedly emphasized in cases of both commercial and residential neighborhood streets in Vietnam [30] and Latin America [15]. Finally, the importance of aesthetic elements, such as the ambiance of historic streets, landscape design, and color perception, is confirmed in the case studies from the USA [27], France [24], and China [31].

In addition, it is recognized that the vitality of the street space can be influenced by the transitional space between streets and buildings. The permeability of the transitional space and the ground paving are thought to attract visitors’ attention and influence their walking-stopping behaviors [16,32]. The openness and transparency of buildings [33,37], the imageability and complexity [26], coherence, legibility, and mystery of building façades can also contribute to the vibrancy of the street [28].

2.2. Tourists’ Walking Behaviors in Historic Spaces

Compared with commercial streets, a particular function of historic streets is cultural preservation, which is closely linked with unique architectural features and cultural elements. These culture-related visual preferences may exert differentiated effects on visitors’ walking behaviors, which have not yet been sufficiently explored.

Several studies have investigated visitors’ walking-stopping behaviors from a spatial perspective. Direct observation was often adopted as a standard tool to register behaviors [37,38,39,40]. Daniel observed and recorded the duration and location of tourists’ stays in Oaxaca’s city center [41]. Zheng et al. determined tourists’ spatial preferences by their time at the sites [42], and Sari et al. examined millennials’ characteristics, intentions, and motivations for traveling in historic tourist areas by observing them [43]. In addition, group and individual measures, such as tourist flow, walking time and speed, walking continuity, and the social interactions and business concomitant behaviors that arise during walking, have also been investigated by scholars [26,44,45,46,47,48]. Studies have also distinguished utilitarian and hedonic walking and compared the consequential walking preferences [1,10,28,49]. In the era of big data and social media, ICT and location-based service (LBS) data have been applied to analyze tourists’ spatial preferences in historic spaces. Maeda extracted the location of tourist destinations from Twitter and Foursquare and made decision trees to determine what combination of characteristics is essential to attract foreign and domestic tourists [50,51]. Li et al. used geotagged photos to estimate tourist trajectories [52]. Liu et al. used GPS to record the patterns of tourists’ spatio-temporal movements [53]. Van der Zee et al. extracted comments on TripAdvisor to describe tourists’ mobile networks using a social network analysis approach [54]. Most of these studies emphasize visitors’ spatial preferences and attempt to create a typology [55,56].

Secondly, abundant studies have focused on tourists’ visual and acoustic preferences toward historic elements. Su obtained tourists’ image perceptions of historic districts through questionnaires and semi-structured interviews to understand the cognitive images of different groups [57]. Deghati Najd et al. implemented a visual preference survey of international tourists based on photographing to evaluate their preferences in different scenes [58]. Montazerolhodjah recorded different sounds heard by tourists in several squares in the historic city of Yazd and determined their relationship with comfort level [59]. Studies have also used Wejchert’s impression curve to assess the visual intensity [25] and simulation tests to model the psychological state during walking [60]. The visual landscape method has also been applied to calculate landscape viewability [61]. In addition, some scholars have used deep learning techniques for image recognition of built environment elements to obtain large-scale measurements of the walkability of streets [17,62] and visitors’ visual preferences [63]. However, it is necessary to note that micro-scale features of the built environment mainly impact walkers through “visual attraction along the way” [19] and “visual interest and stimulation” [20]. Gehl pointed out that pedestrians’ eyes are generally shifted downward by about 10°, so they usually gaze at the bottom of buildings [14]. Features of the built environment can visually attract visitors and induce spontaneous walking activities, including viewing and lingering. In addition, Alexander [64] argued that the scale of space, the relationship of space (including spatial transitions, penetration, and extension), and the boundaries of space largely determine the main qualities of human visual perception, which support the study of visitors’ visual attraction to the micro-scale elements of the built environment. Therefore, street spaces and the ground floor of buildings are recognized as the environmental objects pedestrians focus on [14]. However, the deep learning approach is limited by recognition accuracy and the number of recognizable elements.

One research gap we noticed when analyzing the literature is that the existing studies do not yet completely recognize the tourists’ visual preferences in the micro-scale built environment during walking, and culture-related determinants are largely undermined. This is partially because the micro-scale built environment (MiBE) system has not been fully established, and the relationships between MiBE and walking-stopping behaviors have not been clearly observed and consequently assessed. We believe it is of great importance to synthesize these scattered micro-scale variables to develop a holistic picture to understand those relationships, and incorporating culture-related factors is far more necessary for the future preservation and vitality enhancement of historic streets.

3. Research Design

3.1. Study Area

Our research was conducted in Wudaoying Hutong, a typical form of historic street in the old city of Beijing. In the Yuan and Ming Dynasties, the ancient city of Beijing was divided into several rectangular residential blocks by intersecting main roads and secondary roads [65]. There is a type of representative Chinese traditional courtyard called siheyuan, a rectangular enclosure surrounded by houses on all four sides of a square, located in these residential blocks [66]. Hutongs, in this way, are defined as narrow pathways or alleys joining two rows of siheyuans of different sizes. The historical and cultural values of Hutongs as a typical traditional neighborhood type was not fully recognized in the 1990s. With a vast number of Hutongs being demolished due to years of rapid urban development, their outstanding features and embedded cultural values started to draw public attention. The Hutong tourism has largely been boosted by a proliferation of bars, restaurants, boutiques, and souvenir shops, attracting numerous international and domestic tourists [65].

Located in the fringes of the old city of Beijing and surrounded by numerous examples of historic architecture (e.g., the Lama Temple and Confucius Temple), Wudaoying Hutong links to Lama Temple Street to the east and Anding Gate Inner Street to the west, with a total length of 632 m and a width of 6 m. In this study, Wudaoying Hutong was selected as our study site for the following reasons (Figure 1).

First, Wudaoying Hutong enjoys a long history, and a large number of traditional historic buildings are very well preserved here accordingly. Wudaoying Hutong was initially built in the Ming Dynasty as a military camp for the northern city walls of the old city of Beijing. As groups of civilian craftsmen started to come and settle down here since the Qing Dynasty, the north side of this alley was mainly occupied for military use while the south side had a residential function. Second, abundant MiBE elements have been detected here, especially those with distinctive Chinese traditional and Western modern architectural styles. In 2006, the first group of foreign entrepreneurs entered this area, and a considerable proportion of traditional siheyuans have been gradually transformed into modern restaurants and bars. After years of urban renewal and regeneration, Wuhaoying Hutong has become increasingly attractive to tourists for its multicultural features, including Tibetan culture, Western bar culture, and even exotic customs such as Japanese izakaya [67]. Last but not least, this historic alley is very suitable for assessing tourists’ recreational walking activities, for the majority of pedestrians encountered in this Hutong are non-local visitors. At the same time, local residents and attendants only account for a small proportion of the total.

3.2. Methods

3.2.1. MiBE Variable System

The MiBE variable system constructed in this study is mainly based on the characteristics of pedestrians’ visual perception and lines of sight, ranging from the middle street spaces to buildings on both sides [19,20]. This MiBE system consists of six groups of micro-scale aspects: the scale of streets and buildings, transitional street spaces, ground floor features along the street, building façade features, building colors, and materials, as well as commodity displays. In the horizontal direction, these six groups of features include streets, transitional spaces between streets and buildings, exterior interfaces of buildings, interiors of buildings along with the lines of sight of walking tourists. In the vertical direction, they include elements of the built environment extending from the standard line of sight to the non-standard line of sight of walkers (Figure 2).

The six types of micro-scale elements in the MiBE variable system can be further divided into 12 subtypes and 62 variables, as listed in

Table 2. Micro-scale built environment variable system and its value assignment standards.

Types	Subtypes	Variables	Value Assignment
			0	1	Absolute Value
Scale of streets and buildings	Scale of streets	Street width			From map data
		Ratio of street height to width			Average height/width of both sides of streets
	Scale of buildings	Building width			From map data
		Building height			3.5 m × building floors
Street transitional spaces	Transitional space perception	Floor covering	No	Yes
		Floor uplift	No	Yes
		Space separation	No	Yes
		Setback façade	No	Yes
		Top surface epitaxy	No	Yes
	Street furnitures	Lamp	No	Yes
		Bench	No	Yes
		Dinner table	No	Yes
		Sunshade	No	Yes
		Outside plant	No	Yes
Ground floor features along the streets	Permeability	Deep bottom	No	Yes
		Longitudinal intersection road	No	Yes
	Transparency	Ordinary window	No	Yes
		Trade display window	No	Yes
		Bay-window	No	Yes
		French window	No	Yes
		Glass door	No	Yes
	Tortuosity	Straight ground floor facade	No	Yes
		Zigzag ground floor facade	No	Yes
Building facade features	Chinese architecture features	Shingle slope roof	No	Yes
		Chinese wooden door	No	Yes
		Chinese wooden window	No	Yes
		Stone drum/stone column/stone lion	No	Yes
		Chinese lantern	No	Yes
		Chinese carve patterns	No	Yes
		Coloured drawing	No	Yes
		Door studs/knockers	No	Yes
	Modern architecture features	High door and window	No	Yes
		Straight cornice	No	Yes
		Non openable window	No	Yes
		Entrance and exit without elevation difference	No	Yes
		Smooth architectural slayer	No	Yes
		Simple facade architrave	No	Yes
	Other detail features	Roof terrace	No	Yes
		balcony	No	Yes
		Shop signboard	No	Yes
Building colors and materials	Building colors	Color purity	No	Yes
		Color diversity	No	Yes
		Self-color contrast	Weak	Strong
		Environmental color contrast	Weak	Strong
		RGB Mean RGB			From image data
		RGB SD RGB			From image data
		Lightness Mean			From image data
		Lightness SD			From image data
	Building materials	Glass material	No	Yes
		Brick material	No	Yes
		Wood material	No	Yes
		Concrete material	No	Yes
		Metal material	No	Yes
		Stone material	No	Yes
		Coating material	No	Yes
		Material contrast	Weak	Strong
		Material diversity			The total number of materials contained in the building facade
Commodity displays	-	Artwork	No	Yes
		Clothes and Accessories	No	Yes
		food	No	Yes
		animal	No	Yes
		Inside plant	No	Yes

. The measured values are used for the average RGB value and standard deviation, average lightness value and standard deviation in the street and building scale, and building color. The data for the scale of streets and buildings was also acquired from maps. In addition, the indicator of material diversity is measured by the total number of materials contained in the building facade. The remaining elements are assigned with 0/1 dummy variables, standing for No/Yes or Weak/Strong.

3.2.2. Walking-Stopping Behaviors

Walking-stopping behaviors refer to the interaction between pedestrians and the built environment. According to Gehl [14], the characteristics of the built environment can attract tourists visually and correspondingly create spontaneous walking-stopping activities, including watching and halting behaviors. As observed halting behaviors can vary in Wudaoying Hutong, we further classified those stopping behaviors into two types, enquiring and photographing, based on the extent to which visitors were attracted and their reactions to those attractions. These three types of walking-stopping behaviors as mentioned above were captured in Wudaoying Hutong and are presented in Figure 3, and their definitions and intensity assignments are listed in

Table 3. Tourists’ walking-stopping behaviors and their assigned intensity in our study.

Walking-Stopping Behaviors		Definition	Intensity Assignment
Watching behaviors		A kind of viewing activity of tourists who are visually attracted by MiBE features but do not change their walking direction or show other reactions	1
Halting behaviors	Enquiring	A type of halting behavior when tourists stop and stay for a while to examine details of MiBE charateristics	2
	Photographing	A type of stopping behavior when tourits are impressed by certain micro-scale features and further take photos of those details	3

.

3.2.3. Data Collection

The data collection was carried out on 13 and 14 October 2021, with cloudy weather and temperatures between 11 °C and 20 °C. In order to avoid the busiest commuting time on weekdays, all the data was collected between 10 am and 6 pm. Four well-trained investigators were recruited for the data collection process.

To record the micro-scale elements of the built environment on this historic street, the northern and southern facades of Wudaoying Hutong were photographed panoramically. These facades were later divided into 144 relatively independent building units to conduct further spatial analysis (Figure 4).

The walking behaviors of 143 tourists and the relevant MiBE features that attracted their attention were recorded accordingly. Four investigators were required to track tourists from the eastern or western ends of Wudaoying Hutong and follow them until the other end of the historic alley. During the tracking process, investigators were asked to take photos of the MiBE features that altered the visual attention of the followed tourists or made them stop for a while. The geographical locations of those attractions were recorded on the “six feet” app (http://www.foooooot.com, access date: 16 October 2021), an outdoor recording application for mobile phones to keep track of traveling trails and photos taken along routes, with spatial and temporal data (Figure 5). If the tracked tourists did not walk through the entire street or only passed through the Wudaoying Hutong without sufficient interaction with the surrounding built environment, they were no longer regarded as valid research objects. Children under 14 years old were not included in this survey for they are short in height with a relatively low sight perspective which is much lower than adults. If the tourists were attracted by items other than the built environment itself, such as their mobile phones or clouds in the sky, the corresponding pictures taken by investigators were removed from the dataset. In this way, 109 out of 143 tourists were tracked validly, and 1431 out of 1845 pictures with micro-scale features of the built environment were selected as valid pictures.

3.2.4. Data Processing

The 62 MiBE elements manually extracted from 1431 photos taken by the investigators were later allocated to the 144 units of the building facades according to their geographical location. Then the descriptive statistics were displayed by accumulated assignment of the MiBE features counted by each unit.

Exploratory factor analysis was used in this research to explore the internal relationships among MiBE variables by performing the standard procedure of principal component analysis (PCA). After the irrelevant variables were eliminated at the first step, four primary influencing components by correlation analysis and factor rotation were then determined, all performed in SPSS 22.0 (IBM Corp., Armonk, NY, USA).

Multiple linear regression (MLR) analysis was also conducted to establish a mathematical correlation between the walking activities of the tourists and MiBE features, using the following equation:

$$Y ={ \mathrm{b}}_0+{\mathrm{b}}_1\ast X_1+{\mathrm{b}}_2\ast X_2+{\mathrm{b}}_3\ast X_3+{\mathrm{b}}_4\ast X_4$$

(1)

where $Y$ refers to the intensity of tourists’ walking-stopping behaviors and ${\mathrm{X}}_{\mathrm{i}}$ represents a micro-scale component recognized by the PCA in the previous step. The MLR analysis was also conducted in SPSS 22.0.

4. Results

4.1. Sample Descriptions

Among the 109 tourists, the percentages of males and females (51.4% and 48.6%) were very close. There were 95 Chinese and 14 foreigners, accounting for 87.2% and 12.8%, respectively. A majority of our respondents (n = 87, 79.8%) were teenagers or middle-aged people aged 15 to 59 years old, while only a small proportion of elderly people older than 60 (n = 22, 20.2%) were surveyed here. This is roughly consistent with the population proportions of the female and elderly in the seventh Beijing census, who make up 48.9% and 19.6% of the population, respectively.

A total of 1431 walking-stopping behaviors were recorded in our survey, including 891 watching activities, 334 enquiring, and 206 photographing activities. Each tourist produced an average of 8.2 watching, 3.1 enquiring, and 1.93 photographing behaviors, and each building unit held an average of 6.19 watching, 4.64 enquiring and 4.29 photographing activities (

Table 4. The descriptive statistics of the average cumulative intensity of tourists’ walking-stopping behaviors.

Walking Behavior	Min	Max	Mean	SD
Watching	0	24	6.19	4.68
Halting	0	24	4.64	5.61
(Enquiring)
Halting	0	30	4.29	6.03
(Photographing)
Total	144

). The building unit with the highest intensity of tourists’ walking-stopping activities was generated by the Xingnaichuan Bar (translated from the Chinese name ‘星乃川酒吧’), which reached up to 50. This was followed by the Sky Music Box Store (entitled with ‘天空音乐盒’ in Chinese) and the Sirena Cat Lodging Bar, whose intensity rates were 42 and 38, respectively (Figure 6).

The 88,722 MiBE features extracted from the 1431 photos revealed a higher frequency of tourists’ attraction to trade display windows, glass doors, straight ground floor façades, shop signboards, color purity, glass materials, material diversity, with an average frequency above 2.0 and the standard deviation above 3.0. In addition, other variables attracted tourists, with an average value ranging from 1.0 to 2.0 and a standard deviation above 2.0. These variables included floor coverings, floor uplift, setback façades, top surface epitaxy in the transitional space dimension, brick material, material contrast in the transparency dimension, non-openable windows, color diversity and outside plants (see Supplementary Materials—Table S1).

4.2. Identification of MiBE Variables

By principal component analysis, 15 MiBE variables were identified that exert a significant impact on the walking behaviors of tourists, including bay windows, trade display windows, French windows, glass doors, floor coverings, setback façades, floor uplift, Chinese wooden doors, Chinese lanterns, high doors and windows, color diversity, color purity, glass material, and concrete material. The value of Kaiser–Meyer–Olkin statistic was 0.742, indicating the suitability of conducting an exploratory factor analysis. The Bartlett’s test was significant, with a p-value of 0.000, showing that the variables were interactively related. The total variance of each variable interpretation is presented in

Table 5. Total Variance Explained by MiBE.

Component	Initial Eigenvalues			Extraction Sums of Squared Loadings			Rotation Sums of Squared Loadings
	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	5.327	35.513	35.513	5.327	35.513	35.513	4.581	30.543	30.543
2	2.382	15.878	51.390	2.382	15.878	51.390	2.581	17.206	47.749
3	2.117	14.112	65.503	2.117	14.112	65.503	2.195	14.634	62.383
4	1.440	9.601	75.104	1.440	9.601	75.104	1.908	12.721	75.104
5	0.893	5.954	81.058
6	0.735	4.902	85.960
7	0.500	3.336	89.296
8	0.413	2.753	92.049
9	0.305	2.031	94.080
10	0.219	1.461	95.541
11	0.195	1.299	96.840
12	0.171	1.138	97.978
13	0.124	0.824	98.802
14	0.104	0.696	99.499
15	0.075	0.501	100.000

Extraction Method: Principal Component Analysis.

. Based on this, four components were determined as the principal components to represent the 15 filtered variables, the cumulative contribution of which reached 75.104% in total.

The component and factor loadings are presented in

Table 6. Rotated Component Matrix of MiBE.

Elements	Component *
	1	2	3	4
French window	0.897	0.063	0.107	0.132
trade display window	0.886	0.104	0.099	−0.152
glass material	0.853	0.098	0.227	0.051
color purity	0.828	0.102	−0.313	−0.028
glass door	0.779	0.261	0.140	0.259
bay-window	0.739	0.080	0.203	0.215
high door and window	0.542	0.041	0.495	−0.217
floor uplift	0.149	0.904	0.049	0.251
setback façade	0.106	0.908	0.077	0.071
floor covering	0.173	0.862	−0.037	−0.208
concrete material	0.185	−0.097	0.849	−0.065
Chinese wooden door	0.009	0.051	0.678	0.496
color diversity	0.067	0.217	0.668	0.329
Chinese wooden window	−0.006	0.139	0.041	0.893
Chinese lantern	0.168	−0.076	−0.082	0.715

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. *: Rotation converged in five iterations.

. The first component extracted comprised seven factors (French window, trade display window, glass material, color purity, glass door, bay window, high door, and window), with all loadings greater than 0.5. These items were all linked with features in commercial spaces that displayed a certain degree of transparency, so this component was named herein as interface transparency of retail space. Similarly, the second component contained three factors (floor uplift, back façade, and floor covering) with relatively high loadings above 0.8. As these factors all represented certain traits of the transitional space, it was entitled with transitional space between buildings and streets. The third component was named as the contrast of Chinese and Western styles, for mainly constituting three factors exhibiting cultural differences: concrete material, Chinese wooden doors and, color diversity. The last component, herein labelled with traditional Chinese characteristics, carries two variables of distinctive Chinese features, Chinese wooden windows, and Chinese lanterns.

4.3. Relationships between MiBE Variables and Tourists’ Walking Behaviors

A mathematical correlation between MiBE variables and tourists’ walking activities was established by MLR analysis. The model (Table 6) explained 68.2% of the variables. The final equation was $\mathrm{Y} ={ \mathrm{b}}_0+0.703{\ast \mathrm{X}}_1+0.304{\ast \mathrm{X}}_2+0.208{\ast \mathrm{X}}_3+0.196{\ast \mathrm{X}}_4$. This result displayed a significant correlation between the halting behaviors of tourists and four principal components (${\mathrm{X}}_1, {\mathrm{X}}_2, {\mathrm{X}}_3, {\mathrm{X}}_4$) identified in

Table 7. Multiple linear regression results showing the impact of four primary components on tourists’ walking-stopping behaviors.

Dependent Variable	Y	Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	SE	Beta
Constant	b0	5.694	0.234		24.304	0.000
Independent variable	X1 (Principal Components 1)	3.504	0.235	0.703	14.902	0.000
	X2 (Principal Components 2)	1.697	0.235	0.340	7.216	0.000
	X3 (Principal Components 3)	1.037	0.235	0.208	4.410	0.000
	X4 (Principal Components 4)	0.979	0.235	0.196	4.164	0.000
R	R Square	Adjusted R2	R2 Change	F Change	Sig.F Change	df
0.831	0.691	0.682	0.691	77.727	0.000	4

. These four components were selected as independent variables, and the contribution rates were 49.8%, 21.6%, 14.7% and 13.9%, respectively. This result demonstrates that the transparency of the interface had the most significant impact on pedestrians’ halting behaviors in commercial spaces, followed by the transitional space between streets and buildings. Simultaneously, the contrast between Chinese and Western cultural elements and traditional Chinese characteristics also significantly impacted the walking-stopping behaviors of tourists.

5. Discussion

5.1. The Relationships between MiBE Variables and Tourists’ Walking-Stopping Behaviors

One of our novel findings is that the contrast between Chinese and Western styles, usually integrated with rich and diverse architectural colors, has a noticeable impact on tourists’ walking behaviors. In the third factor of the rotated component matrix listed in Table 5, the loadings of concrete material and Chinese wooden doors (0.849 and 0.678, respectively) are the highest of all variables, signifying the importance of integrating Chinese and Western cultural elements to attract tourists. In addition, the loading of color diversity (0.668) is very close to that of Chinese wooden doors (0.678), ranking third among all variables, further indicating the attractiveness to visitors of mixed colors. Figure 7 displays some typical architecture units belonging to this category. For instance, the design of Xingnaichuan Bar (星乃川酒吧) adopted different architectural materials, including grey bricks, metal rims, and glass windows to incorporate some modern elements into the traditional siheyuans style, which was proven to be eye-catching to visitors. Such elements can induce visitors to stop and look more frequently; some even take pictures with them, but the duration is usually short. It does not trigger more prolonged further actions. We suppose this might be due to the fact that Chinese and Western cultural elements are decorated or embedded in the facade, usually without any change to the architectural interface and without creating any interactive spaces to allow for social exchange. Consequently, visitors can get easy access to understand the cultural characteristics of the space through viewing alone.

In addition, our study demonstrated that traditional Chinese elements on historic streets encourage visitors to stop walking. In the fourth primary factor, Chinese wooden doors had the highest loading at 0.893, followed by Chinese lanterns at 0.715. The regression analysis results also show that traditional Chinese architectural elements correlate significantly with tourists’ halting behaviors. Most of the walking-stopping behaviors associated with this element were watching and photographing, and some visitors even entered the interior to explore more closely. Typical architectural units of this type include the Beijing Humanities and Arts Center with the traditional Chinese door, Tangchu (translated from ‘汤厨’ in Chinese), a restaurant decorated with climbing plants and an alley before entering the courtyard, as well as the 10th siheyuan with hanging Chinese lanterns, displayed in Figure 8. It should be noted that, however, for this type of element, visitors’ interaction is even weaker compared with the contrast between Chinese and Western styles. Traditional Chinese architecture is mainly composed of brick walls and essential elements, such as Chinese wooden doors and windows, without modern consumer-oriented components available for display and interaction, such as glass and retreats, which attract visitors to stay longer and thus socialize.

These two findings suggest that visitors are attracted to cultural elements in historic streets, including traditional Chinese features and a mixture of Chinese and Western styles. This innovative discovery is derived from the MiBE index system proposed in our study to which various culture-related microscopic elements are added, including Chinese lanterns, bay windows, Chinese wooden windows, wooden doors, and so on. These findings address the research gap on how cultural features of the built environment can impact tourists’ walking behaviors and provide a scientific reference for planning and designing in the future.

Our study also recognized the transparency of the interface as the primary factor in attracting visitors to spend more time on historic streets. As shown in Figure 9, the floor-to-ceiling windows with metal or wood frames are frequently used by store owners to achieve a certain level of transparency, through which tourists could see bookshelves, porcelain, coffee makers and other indoor decorations and consequently were attracted to take photos or buy souvenirs. The MiBE elements, such as trade display windows, bay windows, high doors, and windows, can also increase this transparency. This finding verified the observations of Chen and Zhao [34], Xu [33], and Do [30] in commercial streets in Shanghai, China, and Da Nang, Vietnam that transparent glass windows can help to extend the stay of tourists and boost local business. Furthermore, our research indicated that different forms of glass windows, such as glass materials of building façades and bay windows, all have similar effects.

Another important finding of this study is that many halting activities and social behaviors occurred in transitional spaces between streets and buildings. These transitional spaces can be created by setting back the façade of the ground floor of the building, distinguishing the space by the height or pavement difference between the street and the building interface, and decorating the space with a variety of street furniture, such as street lamps, wooden benches, and flower-stands, as displayed in Figure 10. The combination of these design elements can easily catch the eye of tourists and provide spaces for social interaction. This discovery verified the research findings of Nagata et al. [17] and Steinmetz-Wood [18] in non-historic commercial streets in Tokyo, Japan, and Montreal and Toronto, Canada, confirming that such transitional spaces are similarly attractive to visitors in all commercial streets. The intersection and collision of traditional and modern features in the same street are essential to promote the vitality of historic streets.

5.2. Implications for the Preservation and Regeneration of Historic Streets

The combination of Chinese and Western cultural elements mentioned above is, to some extent, a mixed product of globalization and cultural alienation. As Friedman pointed out, however, this phenomenon should be understood from an integrated local and global perspective [68]. This alienation process accelerated after the Second World War, especially in the last three decades of the 20th century. Modern Western culture swept through every corner of the world causing widespread changes. As a node in global networks, local societies have been transformed more dramatically than before. However, the world has not been fully homogenized, and traditional cultures have localized the modernity with a certain degree of autonomy. In reality, scholars have come to recognize that “culture is not disappearing” [69] when globalism and localism come across each other. Foreign cultural elements are subject to “dispositional control” [70] by local subjects, resulting in new cultural configurations that are continuous with the original culture. The contrast and fusion of Chinese and Western cultural elements manifest this unique cultural configuration, which is gradually being accepted and welcomed by the public.

Therefore, our research provides suggests four approaches to the preservation and regeneration of historic streets: (1) preserving the traditional elements of architecture, such as Chinese lanterns, painted carvings, and lattice windows, to fully display the traditional culture and features of historic streets; (2) integrating both traditional and modern architecture styles to create a more substantial visual impact on visitors; (3) enhancing the transparency level of stores and shops along the historic street by setting more glass windows in different forms; and (4) creating more transitional spaces through height differences and adding street furniture and other design methods. All these design guidelines can help to increase the vitality of historic streets.

6. Conclusions

Our study constructed a MiBE variable system to capture the micro-scale features of the built environment, and applied it in the Wudaoying historic street. The main conclusions are summarized as follows: (1) Fifteen micro-scale features that can influence walking behaviors were identified, including bay windows, trade display windows, French windows, glass doors, floor coverings, setback façades, uplifted floors, Chinese wooden doors, Chinese lanterns, high doors and windows, color diversity, color purity, glass material, and concrete material; (2) The transparency of interfaces was recognized as a critical factor affecting tourists’ walking behaviors, contributing to 49.8% of the total activities; (3) The transitional spaces between the street and buildings were also found to be influential on walking-stopping activities of visitors, with 21.6% of activities affected; (4) Two culture-related factors including the contrast of Chinese and Western styles and traditional Chinese characteristics, were found to significantly affect visitors’ behaviors, contributing to 14.7% and 13.9%, respectively. These cultural features are not only displayed by traditional Chinese items, such as Chinese lanterns and Chinese wooden doors, but also can be represented by the integration of traditional and modern elements, such as bay windows. Consequently, there are significant implications for urban planners and architecture designers to better preserve the conventional cultural features of historic streets and adequately integrate them with modern materials, patterns, techniques, and other elements.

While appreciating the merits of this study, there are still several limitations deserving of future research. First of all, the sample of tracked visitors in the present work, though sufficient for our research purposes, is relatively small due to the limited time and resources. As all the data about tourists were collected in autumn during temperate weather conditions; the situation might be quite different in summer with extremely high temperatures or in winter with biting wind. Furthermore, the data collection needs to be carried out simultaneously to exclude the effects of weather differences, which requires more time and surveying personnel. Therefore, it is worthwhile conducting further research on comparing walking behaviors in different seasons and across various types of streets in the future. Furthermore, in the field investigations, we intuitively observed differences in visual preferences based on individual characteristics such as gender, nationality and age. However, the primary object of our study is the tourists’ attraction to culture-related and other micro-scale elements of the built environment and, therefore, did not focus on differences due to demographic characteristics. It is undeniable that the differences in visual preferences between males and females, foreign and domestic tourists, and different age groups could be very interesting topics for further research. In addition, all the visitors were approached and tracked by investigators without questionnaire surveys or in-depth follow-up interviews, restricting the richness and depth of the data. Moreover, all the MiBE variables were manually extracted from photos taken by the surveyors, which helps to increase the data accuracy but simultaneously imposes restrictions on the total amount of data. Techniques such as machine learning should be adopted in the future for built environment studies at larger spatial scales.