Vitality-Oriented Commercial Street Design Strategies: A Multi-Dimensional Quantitative Analysis of Chunxi Road, Chengdu, China

Abstract

Research efforts to explain urban vitality encompass accessibility studies, place-based qualitative studies, morphological analysis, and land use studies. While several of these isolated approaches have yielded promising results, integrating these explanatory frameworks into a single model remains underexplored—and this constitutes the core goal of the present research. For the empirical study on the vitality of a commercial district, 13 explanatory factors were identified, with measured pedestrian flow (as a proxy for street vitality) serving as the dependent variable, examined in the Chunxi Road area of central Chengdu. To account for temporal variations in street vitality, pedestrian flow was measured across different times of the day and days of the week. Bivariate analysis and principal components analysis were employed to develop a multivariate regression model, which was further refined into a predictive algorithm tool to quantify the relative contributions of the explanatory factors. The results indicate that accessibility and street image factors each independently explain a large proportion of the variance in pedestrian flow, while public transport topological distance exerts a negative effect. Notably, the combined model exhibits significantly stronger explanatory power than the individual contributions of various factors reported in existing literature. Beyond advancing theoretical understanding of urban vitality, the primary purpose of this study is to utilize street vitality (operationalized via pedestrian flow) as an optimization indicator for commercial street planning and design schemes. The developed predictive algorithm model serves as a practical tool for designers, providing actionable references during the design formulation process, enabling them to assess potential street vitality based on preliminary design parameters and make evidence-based adjustments to enhance the effectiveness of commercial street designs. Additionally, the study findings offer insights for the management of urban commercial areas to further promote urban vitality.

1. Introduction

Jane Jacobs first linked street vitality to “eyes on the street,” arguing that mixed land use and high pedestrian density foster safe, vibrant spaces [1]. William H. Whyte later supplemented this via 1980s New York observations, highlighting microscale factors (sunlight, seating, edge effects) that boost pedestrians’ willingness to stay, bridging Jacobs’ macro analysis with design details [2]. Gehl expanded the concept by emphasizing “life between buildings” and indicators like pedestrian activity frequency, later linking well-designed spaces to psychological restoration [3]. Talen and Lee reinforced this, noting vitality depends on physical arrangements supporting socioeconomic diversity; their Chicago cases added an equity dimension to the research [4]. With the deepening of research, the concept of urban vitality has gradually become enriched. Contemporary academic research widely acknowledges urban vitality as a multi-faceted concept, encapsulating three key dimensions: economic vitality, manifested in commercial prosperity, high transaction volumes, and stable property values; social vitality, demonstrated by frequent social interactions, strong community cohesion, and diverse public activities; and spatial vitality, characterized by efficient urban space utilization, active street frontages, and high accessibility [5,6]. This multi-dimensional perspective offers a comprehensive framework for understanding and measuring urban vitality. However, it also highlights the necessity of identifying key indicators that can integrate these dimensions. A more nuanced understanding of urban vitality requires not only considering these dimensions but also how they interact over time and space [7].

This study takes Chunxi Road as the research object and provides reference significance for the vitality enhancement of many similar types of districts. As an iconic commercial landmark in Chengdu, China, Chunxi Road has a history spanning over a century, evolving from a busy commercial hub in the early 20th century to a modern integrated business district (Figure 1a). Its typicality lies in perfectly blending traditional commerce with modern consumption. As a representative of China’s urban commercial districts, it reflects the transformation of urban consumption patterns and serves as a window to showcase Chengdu’s vitality, playing a pivotal role in driving the city’s commercial economy and shaping its cultural and commercial identity.

1.1. Pedestrian Flow as a Core Indicator of Urban Vitality

Among the numerous indicators used to measure different sides of urban vitality, pedestrian flow stands out as a direct and comprehensive proxy; it reflects the “people-oriented” essence of urban space. From an economic perspective, pedestrian flow is a direct driver of commercial activity. The Arup report Cities Alive: Towards a walking world provides compelling evidence: after the transformation of an idle pedestrian park in Brooklyn, New York, the increase in pedestrian flow led to a 172% surge in sales of surrounding retail stores; in Bristol, UK, pedestrians spend 65% more than drivers in the same area [8]. Similar findings are echoed in studies by Li et al., which show that areas with high pedestrian flow often experience increased commercial lease rates and business growth [5]. These cases confirm that pedestrian flow is a direct driver of commercial consumption [9]. In terms of social vitality, pedestrian flow is closely linked to social interaction and community cohesion. A study in San Francisco found that residents living near low-traffic, walkable streets have more friends and acquaintances than those near busy, car-dominated streets [8]. Walkable environments encourage pedestrians to stay and interact; street benches, small squares, and shopfronts along pedestrian paths become “social nodes,” where casual conversations, community events, and daily interactions occur naturally. Higher pedestrian flow thus translates to more frequent social connections, stronger community identity, and enhanced social vitality [9]. In terms of spatial vitality, a study in Kuwait City found that well-designed pedestrian-friendly streets can increase the overall attractiveness of urban areas, leading to higher land use efficiency [10]. In summary, pedestrian flow integrates the economic, social, and spatial dimensions of urban vitality, making it a reliable and operable core indicator for evaluating urban vitality.

In recent years, many studies have used big data such as mobile phone signaling data [11,12,13] and social media data [14] as indicators of vitality. This method has many advantages, but it may not be the most accurate for small-scale comprehensive commercial blocks like Chunxi Road. First, data such as mobile phone signaling do not represent the number of pedestrians, but rather the total population within a region (including people in office buildings and residential buildings); second, this type of data has a certain degree of error in positioning accuracy, making it more suitable for large-scale macro-level research rather than small-scale block-level research. In this study, the pedestrian flow of each street was counted by multiple professional staff, ensuring the reliability and accuracy of the data.

1.2. Multi-Dimensional Factors Influencing Pedestrian Flow

While pedestrian flow is critical to urban vitality, it is not an inherent attribute of urban space but is shaped by multiple factors, including accessibility, spatial features, diverse functions (in Chunxi Road it mainly refers to diverse commercial activities) and so on (Figure 1b). These factors interact to determine the attractiveness of a space to pedestrians, thereby affecting pedestrian flow [15,16].

1.2.1. Accessibility

Accessibility refers to the ease with which pedestrians can move between different spaces, which depends on the density and connectivity of the pedestrian network. A well-connected pedestrian network—characterized by dense sidewalks, pedestrian crossings, and underpasses/overpasses at busy intersections—reduces walking time and improves safety, thereby increasing pedestrian flow. In contrast, a fragmented pedestrian network creates “pedestrian barriers,” discouraging walking. A case in Copenhagen showed that after the city expanded its pedestrian network by connecting previously isolated streets and adding pedestrian-only zones, pedestrian flow in the central area increased by 25% within a year [8]. Similar results were found in a study by Ahmed et al., which demonstrated that improving road connectivity in urban areas can significantly boost pedestrian flow [17]. This shows that road accessibility directly affects pedestrians’ travel decisions—convenient movement encourages more walking, while barriers reduce pedestrian flow. Public transport accessibility is another sector which serves as a “gateway” for pedestrian flow: convenient public transport (e.g., subways, buses, trams) increases the reach of urban spaces, attracting pedestrians from wider areas [18]. Research has shown that areas within 500 m of subway stations or major bus stops typically have 30–50% higher pedestrian flow than areas further away (TRID-Database, 2022). This is because public transport reduces reliance on private cars, making walking a more feasible first and last mile of travel. Davies demonstrated the major change in the patronage pattern at anchor stores and local streets in Newcastle as the result of a subway opening [19]. Conversely, areas with poor public transport connectivity—such as suburbs with infrequent bus services—often suffer from low pedestrian flow, even if local amenities are available.

The study of accessibility is essentially a study of spatial structure. In the late 1970s, Bill Hillier and his research team first proposed and applied the space syntax theory [20]. It studies spatial layout’s impact on human activities by quantifying spatial connections, focusing on how the topological structure of road networks influences the potential for pedestrian and vehicle movement, helping analyze urban vitality, traffic flow and how to optimize space use [21]. It provides a quantitative research method for urban spatial structure and has been widely applied in urban research and pedestrian research since its inception. Space syntax theory combined with other data including trading tracks [22], bus trips [23], and pedestrian flow [21] can provide solid support for urban upgrading. In this study, space syntax theory was used to evaluate both walking accessibility and public transport accessibility.

1.2.2. Public Space for Pedestrians

The idea that providing public space dedicated to pedestrians would promote pedestrian activity and public life has a long history in European cities but becomes a formal practice in German cities starting in the 1920s [24]. Gehl suggested that space provisions for pedestrians increased the number of people walking, following his observations of the expansion of Copenhagen’s Strøget pedestrian zone [25,26]. And certainly the various features of these public spaces also significantly influence pedestrian flow. Studies have generally upheld the ability of the public to distinguish qualities of environments including maintenance, which means the image of pedestrian areas also have a definite impact on aggregate behavior in public walking environments [27]. In one such study of shopping areas, it was found that the appeal of the area depended strongly on the maintenance of the whole and the appearance of shop fronts, whereas street activities, greenery, coffee shops, and crowds were also significant in the formation of image and intended choice [28]. Some recent studies have even used street view data, more fully demonstrating the impact of these built environment characteristics on pedestrian flow [29,30,31].

To evaluate the special quality, we conducted a Public Space Public Life (PSPL) survey. The PSPL survey is a classic urban research method. It mainly observes and records the number and behaviors of people in public spaces and the physical features of the spaces, to analyze the relationship between public space environment and public life [26]. It aims to find and understand people’s activities and behavior patterns in the public domain, its results presented by quantitative and qualitative analysis to provide supports for urban public spaces construction and remodel that lead to creating high quality of public space for citizens’ use [32]. Over the years, the PSPL survey has become an important method in the research on urban vitality [33,34,35].

1.2.3. Diverse Commercial Activities

The fundamental demand for research on a commercial district’s vitality originates from the pursuit of commercial profits, and commerce constitutes an integral and vital part of urban vitality research [36]. The distribution of pedestrians across a movement system is related to the intensity of commercial exploitation and so the type of commercial activity. Early studies focused on land use and functional diversity. In recent years, with the development of online services, studying diversity using Point of Interest (POI) data is clearly a more accurate method [6,37,38]. However, these studies remain confined to the dimension of diversity. In commercial blocks, commercial intensity is also an important dimension. The vitality-carrying capacity of large shopping malls clearly differs from that of small stores. Therefore, this paper proposes the concept of “street capacity” to describe the commercial intensity of streets. Its calculation method is “commercial area divided by street length”; a larger result indicates that the street has more functional space per unit length.

1.3. Research Objective—To Establish a Quantitative Evaluation Model for Commercial District Vitality

In 2006, Ratti et al. examined whether location-based services could become a powerful tool for urban analysis if the use of aggregated data in cities was deployed [39]. From then on, data sources such as mobile phone signaling data, social media data, POI (Point of Interest) data, and street view data have all become research tools. However, big data tools are more suitable for large-scale research (e.g., city-scale research) and lack sufficient granularity for block-scale research. The selection of research methods should match the research object; when the research object is Chunxi Road, which is a small-scale area, offline field research and counting will be more accurate than online data in quantitative research.

In the previous sections, we have clarified that the factors influencing urban vitality mainly fall into three categories: accessibility, public space, and commercial activities. Current studies still focus primarily on a single dimension, while comprehensive studies that integrate multiple dimensions remain scarce. This study will collect multiple factors across these three dimensions, aiming to establish a comprehensive commercial block vitality evaluation model to reveal the formation mechanism of commercial block vitality. Given the characteristics of commercial streets, this study, on the basis of commercial diversity, has also added an indicator of commercial intensity, aiming to make the research more complete and comprehensive.

2. Methodology and Data

2.1. Pedestrian Flow Statistics

Chunxi Road is located in the central area of Chengdu (Figure 1a). With a history of over 100 years, it is a typical commercial district in modern and contemporary China. Over the years, the block has developed well, with its vitality level steadily increasing, and thus has strong representativeness. This study focuses on the traditional scope of Chunxi Road, namely the area to the east of Beixin Street, to the south of Zongfu Road, to the west of Hongxing Road, to the north of Dongdajie Street, as well as Nanxin Street, Zhongxin Street and their adjacent street-front areas, covering an area of approximately 20 hectares [40]. In this study, Chunxi Road commercial district was divided into 61 street segments, where each segment is defined as the area between two adjacent intersections. We separated the 61 sections into six areas and labeled them from A1 to F11 (Figure 2a).

Ten trained staff members were hired to conduct this survey. All of them were undergraduate students majoring in architecture, and they received relevant training and instructions prior to the statistical work. During counting, they were uniformly equipped with counters: each was assigned to one of the six zones (A–F), while the remaining four served as substitutes and for rotation. For each street segment, the number of people passing through the street cross-section (Figure 2a) within a fixed 2-min window every hour was recorded. For instance, this included the count of people passing through a certain street cross-section during the 10:00–10:02 period. The counting work started at 10:00 and ended at 22:00 (time-based statistics of pedestrian flow can be found in Figure A1). This research was done on 5 and 6 December 2015 (weekends). The subject of this study is the general laws on how planning factors influence street vitality, rather than the impact of specific dates or seasons on vitality. Over the two weekend days, commercial districts reach the peak value of pedestrians [41], making these two days an appropriate survey period. Based on the on-site survey, the total pedestrian flow of a typical weekend day can be calculated, which can indicate the vitality of Chunxi Road commercial district. The results are as shown in Figure 2b (detailed statistics are shown in

Table A1. Street average daily pedestrian flow in Chunxi Road district.

Sections	Pedestrian Flow	Sections	Pedestrian Flow	Sections	Pedestrian
A1	12,720	B10	5790	E1	33,465
A2	6661	C1	100,095	E2	47,535
A3	13,380	C2	33,045	E3	9210
A4	2745	C3	36,840	E4	4425
A5	13,485	C4	114,165	E5	44,190
A6	14,461	C5	173,835	E6	7170
A7	6195	C6	121,470	E7	88,890
A8	4935	C7	49,020	E8	107,820
A9	610	C8	95,940	E9	40,935
A10	16,861	C9	27,750	F1	21,075
A11	6240	D1	63,720	F2	840
B1	1950	D2	9975	F3	21,330
B2	1155	D3	82,695	F4	7455
B3	2655	D4	7965	F5	22,980
B4	1995	D5	88,740	F6	26,625
B5	17,461	D6	94,815	F7	6270
B6	8175	D7	65,130	F8	30,645
B7	3361	D8	75,330	F9	15,990
B8	14,925	D9	111,015	F10	10,185
B9	2220	D10	45	F11	7185

). The focus of this study is the relationship between planning factors and street vitality, and thus it does not take into account the impacts of factors such as dates and weather on vitality. It can be seen that on a typical weekend, the central cross passage within the block has the highest number of pedestrians, while the pedestrian activity level in the surrounding areas gradually decreases; overall, the southern part of the block has higher vitality than the northern part.

2.2. Analysis of the Spatial Structure of Commercial Blocks

During this research, we built a road segment graph of the Chunxi Road commercial district and its surrounding 1 km buffer zone, using space syntax theory and DepthmapX 0.6.0 software [42].

In the space syntax modeling, we used integration (R500) and total length (R500) to indicate the result. “Integration” is a core quantitative indicator with significant meaning. It measures the ease of connection between a space and all other spaces in the system. A higher value indicates that the space is more accessible and has a wider radiation range, which can intuitively reflect the “accessibility” and “centrality” of urban space [43,44]. According to different scopes of study, the integration value also may be different. Integration (R500) means the research scope has a 500 m radius, which is a suitable distance for walking; this factor is used to indicate the walkability of each street. Total length (R500) refers to the total length of all elements within 500 m of a certain element. It showed the road density of certain areas.

Figure 3a shows the main roads had higher integration (R500) than the branch roads. The streets with the highest integration (R500) were in the center of the district, which seemed to be similar in pedestrian flow research and also identified the main cross. All of the streets in the Chunxi Road district had relatively high integration (R500). The results shown in Figure 3b are similar, with but a slight difference. In this study, the main cross in the district still had the longest total length (R500).

Besides the road network structure itself, the planning of public transport stations is another key factor in the spatial structure of the block. In the beginning of this study, we tried to measure the distance between each street and the nearest public transport station. However, we found that the core area of Chunxi Road has the farthest distance from a public transport station, but it had the greatest pedestrian flow. Therefore, the distance from bus stops obviously was not an important factor affecting vitality in this area. On this basis, we proposed another possibility—that is, the topology depth of each street to the nearest public transport station may be an important factor. This idea is a continuation of space syntax theory, which focuses on the connectivity of roads—specifically the number of road turns rather than distance. The indicator of integration is also calculated based on the number of turns. Additionally, the topology depth of a public transport station refers to the number of turns a pedestrian must make when walking from the station to a certain street. People always walk on a minimum turning degree, which may explain why the central area had a large flow of people, whereas areas close by public transport that had many twists still lacked people.

As shown in Figure 4, the eastern part of the Chunxi Road commercial block had better public transport conditions, more bus stations, and a metro station. The west had fewer stations, and the topological depth generally was higher, which meant it was harder to access. From on-site observations, the foot traffic at metro stations was significantly higher than that at bus stops, which reflects the metro’s strong transport capacity. By comparing the pedestrian flow in Figure 2b, a pattern can be identified: streets with poor public transport accessibility consistently had lower pedestrian flow.

2.3. Evaluation of Public Space Features

The methodology of this survey followed the PSPL research method proposed by Gehl [26], and evaluated the landscape quality of 61 street sections of the Chunxi Road. PSPL research mainly relies on field and questionnaire surveys, whereas this study focused on the spatial quality of streets. Eight factors were selected to evaluate spatial quality, among which Paving, Greening, Decoration, and Facilities had a relatively high correlation with spatial image. These ratings are relatively subjective and rely heavily on the professional competence and personal perception of the raters. Width, Sidewalk, and H/W (Height/Width) ratio have a relatively high correlation with spatial form. The scoring criteria for these factors are relatively more objective, as they are derived from experience and research in relevant architectural design fields. Several architecture students conducted this survey and evaluated a total of 61 street sections A1-F11 in the commercial block of Chunxi Road for street environment. The evaluation explanations of these eight factors are shown in

Table 1. Public space features evaluation factors.

Factors	Details
Paving	Scored 1–5 according to the design and quality of pavement
Greening	Scored 1–5 according to greening rate and diversity
Decoration	Scored 1–5 based on decorations like sculptures and building facades
Facilities	Scored 1–5 based on benches, umbrellas, trash bins and so on
Width	<6 m (1–3 points), 6–20 m (3–9 points), >20 m (10 points)
Sidewalk	With Vehicles (1–9 points), Pedestrian only (10 points)
H/W ratio	>4 or <0.5 (1–3 points), 2–4 or 0.5–1 (4–7 points), 1–2 (8–10 points)
Vision	<100 m (1–3 points), 100–500 m (4–9 points), >500 m (10 points)

H/W ratio is height/width.

.

The results of the PSPL survey (Figure 5a,b) indicated that four factors, Paving, Decoration, Facilities and Sidewalk, have relatively similar score distributions: all have higher scores in the central cross area and lower scores in the surrounding areas, while the score distributions of the other four factors do not show obvious commonalities. In the survey, multiple indicators yielded similar results, indicating that there may be some commonality among these factors. In the subsequent data processing, it is necessary to use appropriate standardization methods to process them.

2.4. Measure of Commercial Business

A commercial district must contain several commercial activities. The data in this study was collected in 2015. At that time, online POI data was not well-developed, so the vast majority of statistics on business formats were completed through on-site surveys. In addition, this study did not include temporary stalls in its statistics. This is because the Chunxi Road block had strict management, making the proportion of temporary stalls extremely low; moreover, during the on-site survey, the concept of pop-up stores was almost non-existent in Chunxi Road. All commercial activities were divided into three types: Retail, Entertainment, and Catering. Then, these three apartments could be further divided into 20 uses. We used the initial word “Retail,” “Entertainment,” or “Catering” and added a number to represent the specific use or function—for example, R1 represents clothing (

Table 2. Classification of commercial formats in Chunxi Road district.

Apartment	Uses	Apartment	Uses
Retail (R)	Clothing (R1)	Entertainment (E)	KTV (E1)
	Suitcase and Bags (R2)		internet café (E2)
	Cosmetics (R3)		Party (E3)
	Jewelry (R4)		Sports (E4)
	Digital products (R5)		Massage (E5)
	Convenience store (R6)		Salon (E6)
	Local Specialty store (R7)		Cinema (E7)
Catering (C)	Chinese fast food (C1)
	Chinese meal (C2)
	Western fast food (C3)
	Western meal (C4)
	Coffee & Drink (C5)
	Tea house (C6)

).

After statistical and visualization processing, the commercial formats distribution in the Chunxi Road district is shown in Figure 6a (detailed distribution is shown in

Table A2. Street commercial diversity analysis in Chunxi Road district.

Sections	Length (m)	Commercial Formats	Total
A1, A3, A5	156	R1, R4, R6, C1, C2	5
A11	62	C1, C2	2
A2	61	L2, C1	2
A4	63	R6, C1, C2	3
A6, A10	113	R1, R6, L1, L2, L5, C1, C2, C5	8
A7, A8	148	R1, C1	2
A9	100	None	0
B1, B3, B4,	133	R1, R6	2
B10	90	C1	1
B2	80	None	0
B5, C2	90	R1, R2, R4, R6, L4, L5, C2, C4, C5, C6	10
B6	77	C1, C2	2
B7	103	R1, R6, C1	3
B8	89	R3, R4, R6, R7, C2, C5	6
B9	64	None	0
C1	130	R1, R2, R4, R7, C3, C4, C5	7
C3	130	R1, R2, R3, R4, R5, L6, C1, C2, C3, C4, C5	11
C4, C5	164	R1, R2, R3, R4, L4, L5, L6, L7, C2, C4, C5, C6	12
C6, D7	79	R1, R2, R3, R4, L5, L6	6
C7	140	R1, R2, R3, R4, R5, R7, L1, L6, C1, C2, C3, C4, C5	13
C8	114	R1, R2, R3, R4, R5, R7, L3, C1, C2, C3, C4, C5	12
C9, D8	78	R1, R2, R3, R4, L5, L6	6
D1, D3	132	R1, R2, R3, R4, C1, C3, C4	7
D2, D10, D11, F2	130	L5, C1, C5	3
D4	112	R1, L6, C5	3
D5, D6	167	R1, R2, R3, R4, R7, C2, C3, C4, C5, C6	10
D9	108	R1, R2, C2, C3, C5	5
E1	86	R6, R7, L7, C1, C2, C4, C5	7
E2, E5	156	R7, L7, C1, C2, C5	5
E3	139	None	0
E4	114	None	0
E6	83	None	0
E7, E8	167	R1, R2, R3, R4, R5, R7, L6, C1, C2, C3, C4, C5	12
E9	71	R1, R2, C3, C5	4
F1, F3	97	R1, R2, R4, R6, C1, C5	6
F11	163	R4, R6	2
F4, F7	170	R1, R6, L5, C1, C2, C5, C6	7
F5, F6, F8	99	R1, R2, R3, R4, R5, R6, R7, L1, L2, L3, L6, C1, C2, C3, C4, C5	16
F9, F10	161	R1, R2, R3, R4, R5, R6, R7, L1, L2, L3, L6, C1, C2, C3, C4, C5	16

). It should be noted that the area with the most commercial diversity is located at the corners of the block, rather than in the central area where the block’s vitality is the highest. This phenomenon to a certain extent challenged the traditional theories of diversity and vitality.

Besides commercial diversity, this study has also added the dimension of commercial intensity and conducted statistics on the commercial area of streets. We used satellite images and on-site measurements to estimate the commercial area (including the multi-story areas of street-front shops if they have more than one floor) on both sides of each street in the Chunxi Road district, and divided the building area by the street length to obtain a new indicator: commercial capacity.

$$C=\mathrm{S}/\mathrm{L},$$

(1)

where C is the commercial capacity, $\mathrm{S}$ is the commercial area of a street, and $\mathrm{L}$ is the street length of a street. The statistical results of commercial capacity are as follows. It can be seen in Figure 6b that the street capacity in the central area of the block is relatively high, but the area with the highest capacity is instead located at the block’s boundary (detailed commercial capacity statistics are shown in

Table A3. Street commercial capacity in Chunxi Road district.

Sections	Capacity (m2/m)	Sections	Capacity (m2/m)
A1, A3, A5	5.13	C8	397.46
A11	1.77	C9, D8	48.72
A2	5.83	D1, D3	222.65
A4	2.70	D2, D10, D11, F2	4.46
A6, A10	25.49	D4	3.57
A7, A8	4.46	D5, D6	74.91
A9	0.00	D9	31.48
B1, B3, B4	2.03	E1	47.21
B10	2.56	E2, E5	30.64
B2	0.00	E3	0.00
B5, C2	535.00	E4	0.00
B6	1.95	E6	0.00
B7	0.68	E7, E8	282.28
B8	21.12	E9	21.83
B9	0.00	F1, F3	14.23
C1	226.00	F11	2.58
C3	159.85	F4, F7	17.29
C4, C5	218.61	F5, F6, F8	153.74
C6, D7	45.57	F9, F10	81.88
C7	331.50	Average	85.45

).

2.5. Establishment of the Vitality Model in Chunxi Road

After the collection of the various data mentioned above is completed, it is necessary to select appropriate models to analyze the data and build a model. Regarding the variables studied in this paper, correlation analysis was used for data cleaning, with the primary step of excluding factors that are not correlated with pedestrian flow. Due to the large number of factors studied in this paper, the internal connections between these factors may affect the accuracy of modeling. Therefore, after data cleaning, data standardization should be conducted. In this study, we adopted the principal component extraction (PCA) method to convert multiple factors into a small number of key intrinsic factors. Principal components have superior performance compared with an original variable. PCA can avoid the collinearity among factors, so as to simplify the model and improve its accuracy [45].

Multiple linear regression offers key advantages for data analysis. First, it quantifies the combined and individual effects of multiple independent variables on a dependent variable, clarifying how each factor influences outcomes. Second, it enables predictive modeling—using known independent variable values to forecast dependent variable trends, aiding evidence-based decisions. Third, it is computationally feasible and interpretable, with clear metrics (e.g., regression coefficients, R-squared) to assess model fit and result reliability, making it widely applicable in social sciences, urban studies, and other fields. After the cleaned and standardized data undergoes multiple linear regression modeling, it will be able to reveal the generation mechanism of pedestrian flow, which indicates the street vitality.

3. Result

3.1. Correlation Analysis Between Factors and Vitality

We entered the daily pedestrian flow into the database using bivariate correlation in IBM SPSS Statistics 31.0.1 to analyze the correlation between variables in pairs. Low correlation indicates that the factors have a weak relationship with vitality, and thus such factors can be excluded to simplify subsequent analysis. Therefore, we excluded greening, whose correlation was just slightly above 0.2. It is worth noting that these three factors—Integration (R500), Total Length (R500) and Sidewalk—have the highest correlation (around 0.7) with pedestrian flow. Correlations of Decorations, Facilities, and Width are also above 0.6. Depth to Bus/Metro shows a negative correlation with vitality, which indicates that the smaller the distance to public transportation, the higher the vitality of the street.

3.2. Principal Component Extraction of Vitality-Related Factors

After excluding irrelevant factors through correlation analysis, 12 factors remained in this study. These factors were still too complex for subsequent research, so principal component extraction was a necessary step. These 12 factors were further divided into three groups for principal component extraction (with grouping based on

Table 3. Factors’ correlation with pedestrian flow.

Aspects	Factors	Correlation	Sig.
Street Accessibility	Integration (R500)	0.703 **	0.000
	Total Length (R500)	0.677 **	0.000
	Depth to Bus/Metro	−0.495 **	0.000
Public Space Features	Paving	0.551 **	0.000
	Greening	0.208 *	0.033
	Decorations	0.632 **	0.000
	Facilities	0.630 **	0.000
	Width	0.610 **	0.000
	Sidewalk	0.683 **	0.000
	Height-width ratio	0.554 **	0.000
	Vision	0.517 **	0.000
Commerce	Capacity	0.478 **	0.000
	Mixed Uses	0.521 **	0.000

Note: * $p<0.05$, ** $p<0.01$, N = 61.

), and their extraction processes are shown in

Table 4. Total Variance Explained of street accessibility factors.

Component	Initial Eigenvalues			Rotation Sums of Squared Loadings
	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	2.110	70.343	70.343	1.590	53.011	53.011
2	0.679	22.633	92.977	1.199	39.966	92.977
3 (Excluded)	0.211	7.023	100.000

Note: Extraction Method: Principal Component Analysis. Kaiser–Meyer–Olkin measure of sampling adequacy: 0.566; Rotation method: varimax with Kaiser normalization.

,

Table 5. Total Variance Explained of public space features factors.

Component	Initial Eigenvalues			Rotation Sums of Squared Loadings
	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	5.103	72.898	72.898	3.502	50.028	50.028
2	1.109	15.849	88.748	2.710	38.720	88.748
3 (Excluded)	0.278	3.977	92.724
4 (Excluded)	0.206	2.936	95.660
5 (Excluded)	0.139	1.983	97.643
6 (Excluded)	0.095	1.358	99.002
7 (Excluded)	0.070	0.998	100.000

Note: Extraction Method: Principal Component Analysis. Kaiser–Meyer–Olkin measure of sampling adequacy: 0.856; Rotation method: varimax with Kaiser normalization.

and

Table 6. Total Variance Explained of commerce factors.

Component	Initial Eigenvalues			Rotation Sums of Squared Loadings
	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	1.624	81.211	81.211	1.624	81.211	81.211
2 (Excluded)	0.376	18.789	100.000

Note: Extraction Method: Principal Component Analysis; Kaiser–Meyer–Olkin measure of sampling adequacy: 0.699; Rotation method: varimax with Kaiser normalization.

. It should be noted that the principal components extracted in this study have eigenvalues greater than 1. However, in the extraction of road network factors (Table 4), if Principal Component 2 were excluded, Principal Component 1 alone could only explain 70.34% of the cases, which is relatively low (usually higher than 80%). In contrast, when Principal Component 2 is included, the explained cases reach 92.98%, resulting in stronger representativeness. Additionally, the Rotation Sums of Squared Loading for Principal Component 2 reaches 39.97%, which is close to the 53.01% of Principal Component 1, indicating strong representativeness. Therefore, Principal Component 2 in road network factors was retained.

Through the above process of PCA extraction, we extracted five principal components from the 12 factors above that had potential influence on street vitality. These five principal components eliminate the potential collinearity that may exist among the previous 12 factors. The result is shown in

Table 7. Factors’ correlation with principal components.

Aspects	Street Accessibility		Public Space Features		Commerce
Principal Components	A (70.3%)	B (22.6%)	C (72.9%)	D (15.8%)	E (72.5%)
Integration (R500)	0.793	−0.499
Total Length (R500)	0.961	−0.106
Depth to Bus/Metro	−0.196	0.969
Paving			0.925	0.195
Decoration			0.888	0.335
Facilities			0.887	0.317
Sidewalk			0.849	0.438
Width			0.423	0.852
Height-width ratio			0.378	0.810
Vision			0.169	0.941
Street Capacity					0.885
Mixed Uses					0.829

Note: Extraction method: principal component analysis; N = 61.

. Component A is mostly correlated with Total Length (R500) and Integration (R500); the correlation is 0.7793 and 0.961, which means it can represent the spatial structure of the street—strongly connected with accessibility—while Component B represented pubic transport accessibility with a correlation at 0.969. Component C is closely connected with Paving, Decoration, Facilities and Sidewalk, which can represent street image to a large extent. Component D is closely connected with Width, Height-width ratio and Vision, which can represent spatial form, Component E is connected with Street Capacity and Mixed uses, so it represented commercial performance. Compared with the previous factors, these five principal components have stronger generalizability and independence. Therefore, the subsequent analysis can be more accurate and more concise.

3.3. Multiple Linear Regression

This study adopted the stepwise method for regression modeling, which involves gradually incorporating independent variables with significant explanatory power for “daily pedestrian flow” through statistical tests. Specifically, only one most significant variable is introduced at a time until adding new variables no longer significantly improves the model. The construction of the three models are shown in

Table 8. Regression models summary.

Model	R	R Square	Adjusted R Square	Std. Error of the Estimate	Change Statistics					Durbin–Watson
					R Square Change	F Change	df1	df2	Sig. F Change
1	0.688 a	0.473	0.463	29,644.306	0.473	49.356	1	55	<0.001
2	0.796 b	0.633	0.620	24,950.913	0.160	23.638	1	54	<0.001
3	0.849 c	0.721	0.705	21,961.727	0.088	16.700	1	53	<0.001	1.718

Note: a. Predictors: (Constant), Principal Component A; b. Predictors: (Constant), Principal Component A, Principal Component C; c. Predictors: (Constant), Principal Component A, Principal Component C, Principal Component B; Dependent Variable: Pedestrian Flow.

. Model 1 includes only “Principal Component A”; Model 2 adds “Principal Component C” on its basis; and Model 3 further incorporates “Principal Component B”.

Overall, as independent variables are gradually included, the models’ explanatory power for the dependent variable continues to be enhanced. The correlation coefficient (R) increases from 0.688 to 0.849, indicating that the linear association between variables and the dependent variable strengthens from “moderate” to “high”. The coefficient of determination (R Square) rises from 47.3% to 72.1%, meaning Model 3 can explain 72.1% of the variation in pedestrian flow. The adjusted R Square (0.705) shows a small gap with R Square, suggesting that all added variables are effective and free of redundancy. The standard error of the estimate decreases from approximately 29,644 to 21,961, reflecting a significant improvement in prediction accuracy. The change statistics indicate that after each addition of variables, the F-test significance corresponding to the R Square change values (16.0%, 8.8%) is less than 0.001, demonstrating that under the screening of the stepwise method, the inclusion of new factors significantly improves the model. Additionally, Model 3 has a Durbin–Watson value of 1.718, which falls within the range of 1.5–2.5, indicating no significant autocorrelation in residuals and satisfying the independence assumption. In summary, Model 3, which includes three factors selected through the stepwise method, is the optimal model, as it can explain pedestrian flow changes more accurately.

Table 9. Coefficients of regression model 3.

Model	Unstandardized Coefficients		Standardized Coefficients	t	Sig.	Collinearity Statistics
	B	Std. Error	Beta			Tolerance	VIF
(Constant)	34,792.610	2911.654		11.949	<0.001
Principal Component A	19,162.741	3172.980	0.490	6.039	<0.001	0.800	1.250
Principal Component C	17,757.075	3218.510	0.447	5.517	<0.001	0.800	1.250
Principal Component B	−11,811.614	2890.341	−0.296	−4.087	<0.001	1.000	1.000

Note: Dependent Variable: Pedestrian Flow.

presents the regression coefficients and related statistical results of three principal components (Principal Component A, B, and C) on the dependent variable “daily pedestrian flow”. In terms of unstandardized coefficients, the constant term is 34,792.610, representing the baseline pedestrian flow when all three principal components are 0. For every 1-unit increase in Principal Component A, the pedestrian flow increases by an average of 19,162.74; for every 1-unit increase in Principal Component C, the pedestrian flow increases by an average of 17,757.08 (both positive effects). In contrast, for every 1-unit increase in Principal Component B, the pedestrian flow decreases by an average of 11,811.61 (negative effect). The standardized coefficients (Beta values) indicate the order of influence strength: A (0.490) > C (0.447) > B (−0.296). In the significance test, the absolute t-values of the three components are all relatively large (6.039, 5.517, −4.087), and their Sig. values are all less than 0.001, demonstrating that their influences are highly significant. Collinearity statistics show that the tolerance of the three principal components is ≥0.800, and the Variance Inflation Factor (VIF) is ≤1.250, far below 10, indicating no significant multicollinearity and reliable model coefficients. In summary, all three principal components have significant impacts on pedestrian flow, with A and C playing a promoting role and B playing an inhibiting role. Considering that B was closely related to the topological depth of the public transport station, the lower the value indicated the greater the connection to public transport. Therefore, the negative correlation between it and the dependent variable was reasonable. Through the above analysis, the explanatory equation for pedestrian flow in the Chunxi Road can be directly derived:

$$Y=\mathrm{34,792.610}+{\mathrm{19,162.741}X}_1+{\mathrm{17,757.075}X}_2-\mathrm{11,811.614}{X}_3$$

(2)

where Y is the pedestrian flow on a typical weekend day on a street; $X_1$ represents Principal Component A (Accessibility); $X_2$ represents Principal Component C (Street Image); and $X_3$ represents Principal Component B (Public Transport). The value of 34,792.610 is a constant, which can be regarded as the fundamental vitality value of the Chunxi Road commercial block, and its composition can be determined only after further study. This model shows that accessibility is the most important indicator affecting street vitality, followed by street image, and then public transport factors. However, the two principal components representing street spatial form and commercial activities were excluded in the regression analysis, which proves that there is no obvious linear relationship between them and street vitality.

3.4. Curve Analysis of Excluded Components

In the linear regression analysis above, Principal Component D (Spatial form) and Principal Component E (Commerce) were excluded by the model, indicating that their relationship with vitality does not fit a linear relationship. The author further used multiple curve models to fit these two principal components to pedestrian flow in an attempt to identify the regularity of their relationship with street vitality.

Table 10. Curved analysis summary for Principal Component D (Spatial form).

Equation	Model Summary					Parameter Estimates
	R Square	F	df1	df2	Sig.	Constant	b1	b2	b3
Linear	0.176	12.593	1	59	<0.001	33,877.869	16,699.205
Inverse	0.003	0.201	1	59	0.656	34,067.085	56.920
Quadratic	0.301	12.495	2	58	<0.001	48,056.013	18,371.786	−14,414.446
Cubic	0.432	14.427	3	57	<0.001	40,126.420	51,509.477	−4313.733	−17,571.939
Compound	0.371	34.743	1	59	<0.001	14,349.080	2.609
Exponential	0.371	34.743	1	59	<0.001	14,349.080	0.959
Logistic	0.371	34.743	1	59	<0.001	6.969 × 10−5	0.383

Note: The independent variable is Principal Component D (Spatial form); Dependent Variable is Daily Pedestrian Flow.

and Figure 7a analyze the fitting effects of seven curve models with “Principal Component D (Spatial Form)” as the independent variable and “Daily Pedestrian Flow” as the dependent variable. The results show that the linear model has low explanatory power (R Square = 0.176), explaining only 17.6% of the variation in pedestrian flow; the inverse model has extremely weak explanatory power (0.3%) and is not significant; the quadratic model improves explanatory power to 30.1%, showing a certain nonlinear relationship; the compound, exponential, and logistic models have the same explanatory power (37.1%) with good fitting effects; while the cubic model has the highest explanatory power (43.2%) and is overall significant (Sig. < 0.001), capturing a more complex nonlinear relationship. Its parameters indicate that pedestrian flow shows a dynamic trend of “rapid increase—slowing growth—accelerated decline” with changes in the spatial form factor. In summary, the relationship between “Principal Component D (Spatial Form)” and “Daily Pedestrian Flow” is most significantly a cubic nonlinear relationship, and the cubic model is the optimal fitting model.

Table 11. Curved analysis summary for Principal Component E (Commerce).

Equation	Model Summary					Parameter Estimates
	R Square	F	df1	df2	Sig.	Constant	b1	b2	b3
Linear	0.276	22.497	1	59	<0.001	33,877.869	20,919.562
Inverse	0.018	1.056	1	59	0.308	35,843.663	727.132
Quadratic	0.407	19.887	2	58	<0.001	49,879.127	35,521.534	−16,267.946
Cubic	0.413	13.387	3	57	<0.001	47,448.103	39,260.060	−10,770.118	−3371.563
Compound	0.364	33.785	1	59	<0.001	14,349.080	2.587
Exponential	0.364	33.785	1	59	<0.001	14,349.080	0.951
Logistic	0.364	33.785	1	59	<0.001	6.969 × 10−5	0.387

Note: The independent variable is Principal Component E (Commerce); Dependent Variable is Daily Pedestrian Flow.

and Figure 7b presented the fitting results of seven curve models with “Principal Component E (Commerce)” as the independent variable and “average daily total pedestrian flow (as a measure of street vitality)” as the dependent variable. The linear model has moderate explanatory power (R Square = 0.276), accounting for only 27.6% of the variation in pedestrian flow. The inverse model shows extremely low explanatory power (0.018) and is not significant (Sig. = 0.308). The compound, exponential, and logistic models share the same explanatory power (0.364), with better fitting effects than the linear model. Notably, the quadratic function (R Square = 0.407) and the cubic function (R Square = 0.413) have similar explanatory power. Both models are overall significant (Sig. < 0.001), and their common feature is that they reflect the trend that street vitality first increases and then decreases as the commercial factor increases. Specifically, the quadratic function, with its linear term (positive) and quadratic term (negative), directly demonstrates this inverted U-shaped pattern, while the cubic function, despite its additional cubic term, still retains the core trend of initial increase followed by a decrease. In summary, the quadratic and cubic functions, with their comparable explanatory power, both effectively capture the nonlinear relationship between the commercial factor and street vitality, highlighting the key trend of “first increasing, then decreasing.”

Since the overall fitting performance of these curve studies is relatively low, with the highest R Square being only around 0.4, the above analysis can help us understand the relationship between street space, commercial factors, and vitality—that is, vitality generally has a relationship of first increasing and then decreasing with them—but it is not sufficient to form definitive model conclusions.

4. Discussion

Chunxi Road is a commercial district with a historical background and relatively mature development. Its business formats and pedestrian flow are relatively stable; therefore, the vitality analysis of this area can reflect the activity characteristics of people within mature commercial districts. Although the research object is only one district, this study conducted high-precision surveys on pedestrian flow characteristics, commercial formats, and street design quality at the street scale, and established a full-day vitality database for typical dates. The data volume is sufficient to support the establishment of the model. This study first conducted a multi-dimensional quantitative analysis on the vitality generation mechanism of Chunxi Road Commercial district. By constructing a research model incorporating 13 influencing factors, quantitative research confirmed that walking accessibility, street landscape, and public transport are the top three most important aspects influencing street vitality, highlighting the significance of road network design, public transportation planning, and street design for building a pedestrian-prioritized block.

Beyond exploring the vitality mechanism, the core contribution of this study lies in taking street vitality as an optimization indicator for commercial street planning and design schemes, and developing a predictive algorithm model tool to support designers in the scheme formulation process. From the perspective of optimizing the design process, this achievement addresses the traditional limitation where commercial street design often relies on subjective experience rather than quantitative indicators: designers can now use the predictive algorithm to input preliminary design parameters (e.g., planned sidewalk network, number of public transport stops, or landscape configuration) and obtain a quantitative assessment of the potential street vitality. This allows for early identification of design weaknesses (e.g., insufficient walking accessibility leading to low vitality) and iterative adjustments before the formal implementation of the scheme, significantly reducing the risk of post-construction modifications and improving the efficiency of the design workflow.

In terms of application prospects, the tool and indicator system proposed in this study have broad adaptability. While validated using Chunxi Road as a case study, the core logic—linking design elements to vitality via quantitative models—can be adjusted to fit commercial blocks of different scales (e.g., community-level commercial streets or regional commercial hubs) and in different cities. For instance, in the design of new commercial areas, planners can reference the model to prioritize the layout of public transport hubs or optimize street landscape configurations; for the renewal of old commercial blocks, the tool can help assess how to balance vitality enhancement with the preservation of historical features. Additionally, as smart city construction advances, the predictive model can be integrated with real-time urban data (e.g., pedestrian flow, business operation status) to realize dynamic adjustment of design schemes, further expanding its application scenarios.

Naturally, this study also has certain limitations. First, the research data of this study was collected in 2015, which is quite a long time from the present. Second, although this study conducted on-site surveys by trained staff members—a method that ensures greater data accuracy—its data coverage is relatively limited: it only includes data from a regular weekend, without considering the impacts of weather, seasons, and other factors. Therefore, the fundamental purpose of this study is not to comprehensively describe the current vitality of Chunxi Road under various weather conditions and seasons. Instead, it aims to explore the universal inherent laws between urban vitality and various planning factors through this classic case. Such laws are objective, just as the research cases on urban vitality conducted by many pioneering researchers years ago still hold guiding significance to this day. Second, this study attempted to deepen the exploration of how commerce affects vitality. A correlation coefficient of around 0.5 between commercial factors and vitality was identified; however, commercial factors were excluded from the final vitality mechanism model. Vitality and commercial factors generally exhibit a relationship of first increasing and then decreasing. Based on this, a hypothesis is proposed: in the core areas with the highest vitality, excessively high vitality may “crowd out” a certain degree of diversity. This phenomenon could be attributed to factors such as exorbitant operating costs (e.g., rising rents squeezing diverse commercial formats). However, this interaction mechanism still requires in-depth follow-up research. For example, conducting separate studies on businesses with different profitability levels. The same applies to spatial form, another factor excluded from the linear model. In this study, vitality and spatial form also exhibit a relationship of first increasing and then decreasing, which to a certain extent indicates that appropriate space, rather than excessively large or small space, is more conducive to commercial agglomeration. However, their more accurate inherent connection requires further research.

Additionally, the vitality model established in this study includes a constant term alongside various variables. This constant term can be interpreted as the “basic vitality” of the Chunxi Road Commercial District, which may stem from its long-term commercial reputation, geographical location advantages, or cultural attributes. Verifying the formation mechanism of this basic vitality through follow-up studies will enable the predictive tool to better account for the inherent characteristics of different commercial blocks, further improving the accuracy of vitality predictions for design schemes.

5. Conclusions

This research used pedestrian flow as the indicator of street vitality, creating a new multi-dimensional study framework of urban vitality that included three aspects—street accessibility, public space features, and commerce—for promoting the vitality in pedestrian blocks. Through space syntax modeling and PSPL research, we collected data for 13 factors according to these three aspects. Then we used SPSS statistics software to complete a multiple linear regression. The resulting formula showed that walking accessibility, street appearance, and public transport were the top three factors affecting street vitality. This result revealed the importance of planning and design of commercial districts. Moreover, the link to public transport was also important, which supported the idea of transit-oriented development. With this new study framework, we can better understand the basic logic of commercial districts’ planning and design. This framework studied urban vitality according to a multi-dimensional perspective, which more accurately revealed the essence of urban vitality than previous one-sided studies.