How Does the Living Street Environment in the Old Urban Districts Affect Walking Behavior? A General Multi-Factor Framework

Abstract

To strike a trade-off between walking behavior and street resource constraint, extensive research tends to focus on how the urban environment affects walking behavior. However, most of the existing impact measurements focus on the cities in low-latitude temperate environments, which may not truly reflect the situation when assessing high-latitude cities. To address this drawback, in this paper, a general multi-factor framework is introduced to quantify the influence of street-level environmental factors on walking behavior. Specifically, a framework is constructed by comprehensively considering the subjective data and the objective data of Harbin, China, which is mainly composed of multivariate measurement indicators, a multi-source data analysis library, and four-dimensional evaluation paradigm. The results indicate that two main measures can promote the current situation of human-oriented living street environment planning, namely, increasing the distribution of green facilities and life service facilities in the old urban districts living street, and paying attention to the diversity of street greening and street landscape. The proposed framework is conducive to improve the planning status of human-centered street environments and guide the construction of pedestrian-friendly life and healthy streets.

1. Introduction

The World Health Organization predicts that by 2050, two-thirds of the world’s population will live in cities [1], which indicates that urbanization development plays a more vital role in the development of social economy and science. However, there are plenty of challenges with the advancement of urbanization, such as environmental pollution, lack of public space, and low natural coverage. At present, urbanization is more demanding to reflect health concepts. The research on the correlation of city–environment–health has deeply attracted the attention of scholars in various fields. As one of the most accessible forms of physical activity, walking behavior can effectively improve obesity and chronic diseases, which are currently among the biggest issues affecting human health and mortality in the world [2]. At the same time, walking is considered to be a carrier of contact between the body and the world and an effective channel for interaction between physical activity and urban environment [3]. Therefore, considering its influence on the walking behavior in urban environmental construction can not only promote the improvement of urban design and promote the process of urban sustainable development, but also effectively improve health status [4].

Along with the social attention to body health, urban planning is increasingly gravitating towards a street environment which can promote pedestrian activity. In order to scientifically evaluate the promoting effect of environmental characteristics on walking behavior, a series of walking evaluation indexes have been put forward. Such as Walk score^® [5] and Walkability Index [6], which are mainly composed of community level environment factors. Whereas since the limited availability of data, only a small number of street-level environmental factors are involved. Furthermore, walkability index at street level has been shown to be generally effective [7,8]. Alfonzo (2005) introduced evaluation measures of feasibility, accessibility, safety, comfort, and pleasure to indicate the needs of different levels of walking in community and street environments. Note that the street, which is unlike the neighborhood space, is associated with high levels of pedestrian demand [9]. In 1997, Cerveor et al. indicated that the built environment which has an impact on physical activity has 3-Dimensional (3D) characteristics, including density, diversity and design. Inspired by this, Ewing and Cervero improved and added destination accessibility and traffic distance to form 5-Dimensional (5D) elements for evaluating the built environment. Significantly, accessibility represents the distance of residents from the starting point to the critical point, which is an important element in optimizing walking networks in building pedestrian-friendly cities. Research on the evolution of accessibility can reduce inequalities in the distribution of urban infrastructure and services and improve urban resilience [10]. In fact, the built environment of streets has a guiding effect on pedestrian behavior [11]. Therefore, the assessment of walkability from street-level environmental factors is naturally used to describe the influence of comprehensive environment on walking behavior. In addition, density, street connectivity, and land use mixed have been shown to be features of the built environment that were significantly correlated with traffic walking behavior [12]. It is worth emphasizing that many people living in some areas with high density accessibility and road connectivity tend to have higher health levels [13]. Michael Southworth believes that the environmental quality of pedestrian space is the key to promoting walking behavior, and further proposes that walkable street network should meet several design standards such as space safety and connectivity [14]. Bradley Bereitschaft selected streets with different social backgrounds to evaluate walkability, and introduced that transparency, complexity, cleanliness, and openness of street interfaces affect walkability [15]. A single priority scale was created using the analytical hierarchy process, which provides a favorable tool for creating pedestrian-friendly and sustainable cities [16]. As for recreational walking behavior, perceived environmental factors have a greater impact than morphological factors such as accessibility and connectivity, and people living in vegetation-rich communities are more conducive to the occurrence of walking behavior [17]. The aesthetic sense of the built environment and personal satisfaction with the environment have a positive promoting effect on walking behavior [18,19].

However, although many studies have been conducted on the effects of the built environment on walking behavior, not enough attention has been paid to studying the neighborhood environment from a specific regional perspective, especially integrating the pedestrian perspective with the neighborhood environment. The main reasons are as follows: (1) Walkability evaluation of streets can be generally divided into subjective evaluation and objective evaluation, with the former using index system evaluation in the form of questionnaire or perceptive evaluation in the form of interview. Although the proposal and mature application of these methods have laid the foundation for such research, they are difficult to apply to large-scale urban areas because they are often accompanied by cumbersome workload. At the same time, due to the different backgrounds of the interviewees, the results will have a large degree of subjective error, which makes it difficult to conduct further research only relying on the subjective evaluation results lack of universality. (2) The application of new technologies reveals the possibility of measuring street walkability from the city level, which overcomes the limitations of traditional methods. In recent years, big data methods have been explored by a large number of scholars. A street view image is used for semantic segmentation to analyze the environmental elements of streets and their impact on pedestrian behavior and number. Compared with traditional methods, street view image can provide more comprehensive information with lower time cost [20]. It has been pointed out in [21,22] that combining Geographic information data such as point of interest (POI) and open street map with Geographic information system (GIS) to form multi-source datasets can make the evaluation of street space or walkability more comprehensive and accurate and improve the investigation efficiency. However, most of the evaluation of street environmental elements and street walkability based on the new algorithm are single-dimensional data, and the evaluation results are often more biased to objective measurements. In addition, they often ignore the subjective measurement of the street environment from the human perspective, resulting in the street walkability evaluation being limited by one-sided analysis.

To remedy this issue, this paper focuses on the impact of high-latitude lifestyle street environment on pedestrian walking behavior when analyzing the influence of living street environment on pedestrian walking behavior in the old city districts. A multi-source data analysis framework is constructed by integrating social questionnaire survey, geographic information data (POI, road network data), and land use data, aims to evaluate the streets environment and pedestrian walking behavior in the old city districts of Harbin in China, which can not only explore the impact of environmental factors on pedestrian walking time from the street scale, but also provide a framework for assessing the walkability of residential streets in old urban areas of high latitude. Specifically, the traditional 3D evaluation method is further expanded, and destination accessibility reflecting the degree of regional convenience is added on the basis of its measurement factors, forming a 4-Dimensional (4D) framework combining 3D indicators with the characteristics of living street. By doing so, it can not only evaluate the influence of convenience degree of living street in old city districts on pedestrian walking behavior, but it can also help urban planners to further improve the living circle of residents in the old city districts and enhance the convenience of pedestrians in living streets. Compared to the traditional 3D evaluation methods, the constructed 4D evaluation framework can more comprehensively quantify the morphological environment of the living streets and lay the foundation for a quantitative study of the morphological environment of the living streets in old urban districts. The 4D includes density, diversity, design, and destination accessibility. The statistical results indicate that the living street environment in the old city districts has different degrees of significant impact on walking time, among which the destination accessibility and POI density in the streets morphological environment are the main reasons for prolonging the leisure walking behavior of pedestrians, and the landscape degree and the completeness of safety facilities in the streets perception environment are positive factors to guide people to walk.

2. Materials and Methods

In order to comprehensively investigate the influence of street environment on walking behavior, a general multi-element framework for pedestrian walking time assessment in living streets in old city districts is developed.

The framework indicates the research objectives, research variables, research methods and data sources, as shown in Figure 1. More specifically, the framework of the assessment method includes two dimensions of the street environment, walking time and four measures (street morphological environment, street perceived environment, traffic walking behavior, and leisure walking behavior). It should be noted that statistical analysis of density, diversity, design, and destination accessibility of environmental elements was implemented by GIS to avoid errors and complexity caused by manual calculation. Accordingly, factors such as street cleanliness, street environment satisfaction, degree of vegetation, and landscape were analyzed through questionnaire analysis to highlight the influence of pedestrian subjective intentions in the measurement process.

2.1. Area Scope and Object Selection

To comprehensively evaluate the impact of street environment in the old city districts on pedestrian behavior, eight streets in Nangang District, Xiangfang District, Daoli District, and Daowai District of Harbin have been investigated after field investigation and information screening as shown in Figure 2, including Qingbin Street (serial number) and Fanrong Street (serial number 4) in Nangang District, Situ Street (serial number 2) and Junli Street (serial number 1) in Xiangfang District, Heqing Street (serial number 6) and Minan Street (serial number 5) in Daoli District, and Nanwudao Street (serial number 7) and Daoyoufang Street (serial number 8) in Daowai District.

On the one hand, the selected streets have a long construction time and relatively fixed form environment, and the environmental quality and environmental safety level of the streets are in line with the environmental characteristics of the old city district, with the street length ranging from 600 m to 1800 m, as shown in Figure 3. It can be seen that the shortest is Nanwudao Street, which is 680 m long, and the longest is Qingbin Street, which is 1800 m long. These streets cover various areas of Harbin’s old city district, and their long construction time makes them more suitable as representative street environments for each area. It is worth noting that the map in Figure 2 is from Google Earth, where the red dotted line in the outermost circle is the boundary of the main city district, and the dotted line in the innermost circle is the boundary of the old city district. The orange-shaded part in the image is thought to be the Harbin old city area. On the other hand, it is argued that living streets carry the daily public activities of residents, and their use as investigation areas is conducive to a more comprehensive analysis of the essential impact. In conclusion, in order to ensure that the survey area is a living street, the survey takes the center point of the selected street as the circle, the 15-min living circle as the radius, and selects the street with the highest proportion of POI related to the daily life of the residents in the area. In addition, the purpose of preferentially selecting streets with residents with large differences in age and social class is to increase the sample diversity.

2.2. Research Methods and Data Sources

Since walking behavior is affected by both subjective awareness and objective factors, considering the relationship between subjective perception results and objective factors simultaneously is beneficial to effectively quantify data characteristics.

The questionnaire survey is used for subjective research, which mainly includes the basic social demographic information of pedestrians, the elements of street environment perception, the walking time of pedestrians, and the subjective feeling evaluation of pedestrians on street environment. The questionnaire was distributed to the residents or businesses near the selected eight streets with a survey period of one week. A total of 1500 questionnaires were distributed, and 1456 samples were finally obtained. When the incomplete questionnaires were removed, the total valid sample questionnaires included 1389, with an effective recovery rate of $92.6\%$. In order to eliminate the interference of other factors on the questionnaire survey results as much as possible, the survey was conducted in a relatively stable period of climate conditions. Furthermore, accurate data information was guaranteed by manual verification and coordinate verification. After the reliability test of the result data, the overall Cronbach $\alpha$ coefficient of the questionnaire was 0.739, indicating that the quality of the recovered questionnaire data is good.

Objective research utilizes multi-source data for evaluation, measurement, and analysis, including road network data and POI data. By integrating data resources and using GIS to analyze the data, the morphological environment of the street is evaluated. The road network data comes from the open street map, and the POI data comes from the Google Maps, of which there are six major classes, 16 middle classes, and several subclasses. When calculating POI density, facility diversity, and accessibility, the sample network is set as 40 m outward along the center line of the street. In order to avoid sampling points being affected by accidental errors, this study considers the POI quantity and density of each facility within the entire sample street and its buffer zone on both sides, as shown in Figure 4.

Significantly, due to the complexity of the multiple influence factors between the street environment and the walking behavior, a data-based structural test equation is constructed to describe the presence of multiple influence paths. In other words, the real influence factors of street environment on walking behavior can be revealed by testing the relationship between street form environment and walking time as well as the relationship between perceived environment and walking time. Last but not least, a combination of subjective and objective techniques is adopted to measure the relevant variables.

2.3. Analysis of Variable Factors

2.3.1. Analysis of Walking Behavior Variables

The users of living streets are usually residents in nearby communities, whose behaviors and activities revolve around the main body of life. Hence, pedestrian walking behavior can be divided into traffic walking and leisure walking, as shown in

Table 1. Measurement and descriptive statistics of walking behavior variables.

Variable	Explanation (Unit: h)	Sources	Mean Value	Standard Deviation
Traffic walking behaviour	Take times on traffic walking behaviours	Social Survey	2.52	0.29
Leisure walking behaviour	Take times on leisure walking behaviours	Social Survey	2.46	0.41

. The former mainly refers to the walking behavior caused by necessary life events such as commuting to and from work, school and picking up children, while the latter mainly refers to the walking behavior guided by leisure and entertainment life such as exercise walking, shopping, and gathering with friends. It is argued that walking time plays a representative role in the quantitative analysis of walking behavior [23]. Therefore, traffic walking time is taken as the explanatory variable of traffic walking behavior to analyze the correlation between traffic walking and streets environment. Leisure walking time is considered as the explanatory variable of leisure walking behavior to further discuss the correlation between leisure walking behavior and street environment. Note that walking time is measured in hours in the questionnaire.

According to the survey, $52.29\%$ of the respondents spent no more than one hour per day on traffic walking on the streets, and $27.45\%$ spent no more than two hours per day on traffic walking on the streets. Correspondingly, $38.56\%$ of the respondents spend less than one hour walking on the streets every day, and $30.72\%$ spend less than two hours walking on the streets every day. The statistical data demonstrate that the residents in the old city districts of Harbin lack walking behavior in living streets, and the phenomenon of lack of physical activity is common in the old city residents selected for the study.

2.3.2. Analysis of Street Environment Variables

Residential street environment can seriously affect the lives of residents, because it is the guarantee of basic life services and the embodiment of life quality. Therefore, investigating the impact of the built environment on walking behavior in combination with the characteristics of street areas is beneficial for decision makers to customize relevant policies to improve the quality of life of residents. The environmental characteristics of the sample streets selected in Figure 1 are evaluated, and the street environment is divided into the street morphological environmental elements and the street perception environmental elements. The morphological environment is mainly composed of the density of the street environment, the diversity of the street environment, the design of the street environment, and the accessibility of the street environment. The street perceived environment includes the street environmental quality and the street environmental safety, as shown in

Table 2. Measurement and descriptive statistics of street environment variables.

Variable	Explanation	Sources	Mean Value	Standard Deviation
POI Density	Number of POI within the selected street area (Unit: 1000)	Social Survey	0.04	0.06
Land Use Mixedness	Entropy Index $H=-{\sum }_{i=1}^n{S}_i\times ln{S}_i$	Social Survey	2.46	0.41
Design	The ratio of the number of intersections within the selected street to the street area	Map Website	2.25	2.55
Proportion of POI in Traffic Service Facilities	Proportion of POI in Traffic Service Facilities within the selected street area to the total POI in this area	Map Website	6%	2.8%
Proportion of POI in Catering Facilities	Proportion of POI in Catering Facilities within the selected street area to the total POI in this area	Map Website	22%	22.94%
Proportion of POI in Commercial Facilities	Proportion of POI in Commercial Facilities within the selected street area to the total POIin this area	Map Website	35.78%	36.06%
Proportion of POI in Park Green Facilities	Proportion of POI in Park Green Facilities within the selected street area to the total POI in this area	Map Website	1%	0.3%
Proportion of POI in Public Services Facilities	Proportion of POI in Public Services facilities within the selected street area to the total POI in this area	Map Website	17.92%	19.77%
Proportion of POI in Life Services Facilities	Proportion of POI in Life Services facilities within the selected street area to the total POI in this area	Map Website	23.46%	24.53%
Walking Environmental Satisfaction	Very Satisfied = 5 Very Dissatisfied = 1	Social Survey	2.34	0.21
Plant Richness	Very Plentiful = 5 Very Scarce = 1	Social Survey	2.62	0.23
Landscape Richness	Very Plentiful = 5 Very Scarce = 1	Social Survey	2.59	0.24
Street Cleanliness	Very Satisfied = 5 Very Dissatisfied = 1	Social Survey	2.31	0.25
Street Traffic Security	Not Very Worried = 5 Very Worried = 1	Social Survey	2.45	0.21
Street safety facilities complete degree	Very Inomplete = 5 Very Complete = 1	Social Survey	2.94	0.33

.

The street environmental density is a quantitative index that reflects the spatial distribution characteristics of various facilities in the street. Compared with the population density used in the traditional built environment research, POI density can better reflect the correlation degree between the distribution of facilities in the street and the walking time. According to the regional characteristics of Harbin, POI density is used as the measurement factor of street density. The number of relevant POIs in the selected streets is obtained using the mapping application.

Diversity indicates the degree of mixed use of land in living streets. In the living street, the degree of mixed land use represents the proportion of different functional types of land use in a given area that are closely related to the daily activities of pedestrians. As reported in [24], the degree of mixed land use can effectively guide pedestrians to choose non-motorized travel modes. Hence, the degree of mixed land is employed to measure the diversity of the street environment. In this paper, the measurement of land use mixing degree is divided into two categories, namely, the overall measurement and the zoning measurement. The former is usually used to reflect the proportion of various types of land in the overall area, and in practical application, it is more targeted at small and medium-sized areas, such as residential areas and streets. Within the measurement method, the balance index and the entropy index [25] are well-known indicators. The latter is usually utilized to reflect the degree of uniformity of various types of land in the overall area, and can also be used as a method of determining the degree of similarity between two areas, which can be applied to larger areas, such as cities or towns. Specifically, the entropy index [26] is employed to account for the degree of land use mixing within a selected area. The higher the entropy value, the higher the degree of land use mix as well as the more diverse the land use function in the region.

$$1-(\frac{{\sum }_{i=1}{n}_i(n_i-1)}{N(N-1)})$$

(1)

where n represents the number of interest points distributed for each type of facility within a buffer zone set outward along the center line of each sample street; i represents transportation service facilities, catering facilities, commercial facilities, parks and green space facilities, public service facilities, and life service facilities. N represents the total number of interest points distribution of transportation service facilities, catering facilities, commercial facilities, park green space facilities, public service facilities and life service facilities.

The concept of design has been regarded as the design of street network in a certain area since it first entered the field of planning, rather than the traditional design at the urban level [27]. Compared with other types of streets, living streets are more closely related to daily life. Therefore, it is particularly important to explore whether the shape design of living streets has an impact on pedestrian walking behavior. As a measurement method of design, intersection density is widely used in old city districts streets with open and varied forms [28]. Similarly, intersection density has been used to investigate environmental attributes, especially community-level factors, that influence physical activity levels among older African American women [29]. Considering the regional characteristics of Harbin, intersection density is used as a measurement factor designed in the 4D framework, which can be calculated by the ratio obtained between the number of intersections in the region and the area in the region.

Accessibility, which represents the convenience of pedestrians walking from the starting point to the destination, is widely used to explain the interaction between various facilities nodes in the street network system. The recently proposed living circle concept is to make full use of accessibility to optimize the built environment and enhance the quality of life of residents. As explained in [30], the higher travel frequency of residents in communities with better accessibility indicates that there is a certain correlation between accessibility and residents’ travel behavior. In this paper, since the different types of pedestrian walking destinations in the street, accessibility of transportation service facilities (bus stations, subway stations), public service facilities (such as supermarkets, schools, clinics), commercial facilities, catering facilities, and park green space facilities are selected as quantitative factors. Using the POI data obtained from the map website, the ratio of the number of interest points of each facility in the selected street to the number of interest points of all facilities in the area is the destination accessibility. It should be emphasized that the greater the value obtained, the better the accessibility and convenience of the area.

Additionally, street cleanliness, walking environment satisfaction, vegetation richness, and landscape richness are selected as explanatory factors for the environmental quality of the street. Traffic safety and the level of street safety facilities are considered as explanatory variables of street environmental safety. This part of the explanatory variables follows the subjective questionnaire data definition [31,32].

2.3.3. Analysis of Socio-Demographic Characteristics Variables

Generally, it is different to elaborate the influence of street environment on walking behavior based upon only one single independent variable experiment. Hence, we combined social demographic characteristics as the control variable of the experiment to make the study more complete and comprehensive. Since the socio-demographic characteristics of the surveyed individuals will have an impact on the walking behavior and walking experience of pedestrians [33], socio-demographic variables are taken as control variables, including gender, age, education level, and economic income of the sample, as shown in

Table 3. Measurement and descriptive statistics of Sociodemographic characteristics variables.

Variable	Explanation	Sources	Mean Value	Standard Deviation
Age	Ages 10 to 20 = 1, 21 to 30 = 2, 31 to 40 = 3, 41 to 50 = 4, 51 or above = 5	Social Survey	0.77	0.11
Gender	Male = 0, Female = 1	Social Survey	2.94	0.60
Level of education	Primary education = 1, Secondary school education = 2, High school education = 3, Bachelor degree = 4, Graduate degree or higher = 5	Social Survey	3.87	0.66
Income situation	Lower Level = 1 High Level = 5	Social Survey	2.22	0.31

. After statistical analysis of the collected data, it is found that most of the participants in the survey are between 21 and 30 years old, that is, the age of investigators can be considered relatively young. Within these respondents, $61\%$ of them have higher education, $66.83\%$ of them are female and $33.17\%$ of them are male.

3. Result

3.1. Influence of the Living Street Environment on Traffic Walking Time

According to the statistical results of the street morphological environment on traffic walking time (

Table 4. Results of the correlation analysis.

	Leisure Walking Time		Leisure Walking Time
	Model 1	Model 2	Model 3	Model 4
POI	$-0.757{}^{*}$	$-0.752{}^{*}$	$-0.732{}^{*}$	$-0.726{}^{*}$
Land Use Mixedness	$0.853{}^{**}$	$0.849{}^{**}$	0.637	0.582
Design	0.625	0.619	0.459	0.435
Proportion of POI in Traffic Service Facilities	$0.891{}^{**}$	$0.885{}^{**}$	0.654	0.646
Proportion of POI in Catering Facilities	$0.818{}^{**}$	$0.803{}^{**}$	$0.582{}^{*}$	0.569
Proportion of POI in Commercial Facilities	$0.882{}^{**}$	$0.845{}^{**}$	0.529	0.505
Proportion of POI in Park Green Facilities	$0.830{}^{**}$	$0.819{}^{**}$	$0.715{}^{*}$	$0.708{}^{*}$
Proportion of POI in Public services facilities	0.657	0.621	0.629	0.613
Proportion of POI in Life services facilities	0.536	0.513	$0.761{}^{*}$	$0.754{}^{*}$
Walking Environmental Satisfaction	-	0.445	-	-0.658
Plant Richness	-	$-0.809{}^{**}$	-	$-0.835{}^{**}$
Landscape Richness	-	$0.827{}^{**}$	-	$0.874{}^{**}$
Street Cleanliness	-	0.535	-	$0.741{}^{*}$
Street Traffic Security	-	$0.768{}^{*}$	-	0.623
Street safety facilities complete degree	-	$0.773{}^{*}$	-	$0.928{}^{***}$
Age	−0.280	−0.352	$0.853{}^{**}$	$0.846{}^{**}$
Gender	0.571	0.583	0.476	0.458
Level of education	−0.600	−0.568	$-0.860{}^{**}$	$-0.847{}^{**}$
Income situation	−0.465	−0.442	−0.648	−0.625
$R^2$	0.081	0.072	0.089	0.091
Sig	0.000	0.000	0.000	0.000

“$***$”, “$**$”, and “*” are significant at 0.01, 0.05 and 0.1 levels, respectively.

, Model 1), it is believed that there is no significant correlation between traffic walking time and POI density. However, the statistical conclusions obtained in the correlation analysis tables of the street environment and pedestrian walking time indicate that POI density is negatively correlated with transit walking time. The reason for this result may be that the quantity distribution of facilities POI in the living street in old city districts is much higher than that in other types of streets. In fact, when the density of POI is higher than a certain degree, the aggregation of facilities and the mixing of functions will greatly shorten the walking distance of people. As a result, less traffic walking time will be spent near streets with dense distribution of facilities.

The statistical results show that the diversity of the street morphological environment has a significant correlation with traffic walking time at the level of 0.05. Specifically, the degree of land use mixed in the living street in the old city districts has a promotion effect on pedestrian traffic walking time, which is similar to the analysis results of commercial districts in Hangzhou in literature [34]. In other words, the land using mixed land use attributes most strongly associated with walking for transport is a priority. Furthermore, the higher the degree of land used mixed in the living streets of the old city district, the more beneficial it is to pedestrian traffic walking, and the more helpful it is to guide pedestrians to choose a more active and healthy way of travel.

The accessibility of traffic facilities, catering facilities, commercial facilities, and park green space facilities is significantly correlated with the traffic walking time of pedestrians, which indicates that the destination accessibility performance in the old urban district living streets can guide the traffic walking of pedestrians to a certain extent. The higher the destination accessibility of each facility, the more beneficial impacts on the extension of pedestrian traffic walking time.

In addition, the street environment is further improved to analyze the impact of the street environment on pedestrian traffic walking time more comprehensively. Specifically, the perceived environment element of the street is added as shown in Model 2, which can reflect the environmental quality, environmental perception and the satisfaction degree of the street environment. The statistical results show that there is no significant correlation between walking environment satisfaction and cleanliness of streets and the traffic walking time of pedestrians. The reason is that traffic walking is mostly a necessary travel behavior, and less attention is paid to street cleanliness and street environment satisfaction. However, the results show that the street plant richness is significantly negatively correlated with traffic walking behavior, which is contrary to most research conclusions and common cognition. With a wide geographical area in Harbin, the streets with plant richness and high coverage may have a higher quality of life. Residents living in this area usually travel by motor vehicle, which leads to a low traffic walking time and a negative correlation with the street richness. In order to analyze and study the relationship between street safety and traffic walking time, the degree of traffic safety and safety facilities completeness of the street is considered, as shown in Table 4 and Model 2. It can be observed that there is a correlation between street safety and traffic walking time, which indicates that the higher traffic safety and the better safety facilities of the living streets in the old city are, the more conducive to the extension of traffic walking time.

3.2. Influence of the Living Street Environment on Leisure Walking Time

According to the questionnaire survey on pedestrian walking time, $93.8\%$ of pedestrians are willing to walk for leisure at different times on the street. The statistical results of the influence of the street morphological environment on leisure walking time are calculated (Table 4, Model 3), in which the POI density shows a positive correlation with leisure walking time, indicates that the increase in the number of POIs in the living streets of the old city can promote the extension of the leisure walking time. On the contrary, there is no significant correlation between land use mixing degree and leisure walking time, which may be due to the fact that leisure walking behavior is a non-necessary travel behavior, and the correlation between the occurrence of behavior and the mixing degree of functional facilities of the street is weak.

However, unlike the statistical results related to traffic walking time, the accessibility of traffic facilities, catering facilities and commercial facilities in a street has no significant correlation with leisure walking time, but has a significant positive correlation with the accessibility of park green space facilities and living service facilities. Intuitively, appropriately improving the distribution of park green space and facilities related to life services in the living streets of the old city districts can effectively promote leisure walking behavior and extend its time, which is conducive to guiding pedestrians to a healthier leisure way.

In addition, on the basis of the living street morphological environment in the old city districts, the purpose of combining the street perception environment elements is to further study the influence of leisure walking time. As shown in Table 4 and Model 4, plant richness, landscape richness and street cleanliness are significantly correlated with leisure walking time. When city–environment–health is taken as the direction of urban planning, it is considered to be a very effective measure to enhance the richness of street plants and landscape and improve the cleanliness of the street environment. By doing so, it is beneficial to extend the leisure walking time of pedestrians in living streets, and guide pedestrians to a more healthy leisure way. In addition to conducting a questionnaire survey on the perceived environment of the street, street safety is considered as an indispensable element. The results of correlation analysis indicate that the adequacy of street safety facilities is significantly positively correlated with leisure walking time at 0.01 level, but not significantly correlated with traffic safety. The reason for this may be that compared with traffic walking, pedestrians on the living streets in the old city have more time freedom to choose, and there are fewer pedestrians and traffic flowing on the living streets. Therefore, pedestrians in leisure walking are more concerned about whether the safety facilities of the streets are perfect.

4. Conclusions

This paper examines the impact of street environments on pedestrian walking behavior from the perspective of pedestrians. Specifically focusing on Harbin in China as the research subject, we propose a general multi-element framework for assessing pedestrian walking time in living streets within high-latitude old urban districts. Specifically, we establish a multivariate measurement index and a database that incorporates both measured data and questionnaire survey data from multiple sources. Additionally, we construct a 4D measuring framework to study the varying effects of street environments on pedestrian traffic walking behavior and leisure walking behavior in old city districts.

The results show that pedestrians prefer the street environment with better destination accessibility and higher safety when they walk; for example, the street environment with a rich distribution of park green space facilities and a high level of safety facilities. In addition, walking behavior for transportation and leisure purposes is affected by different street environment elements. In particular, POI density of the street, plant richness degree, and landscape degree of the street inhibit the time of pedestrian traffic walking. Interventions to promote traffic walking behaviour should increase the distribution of traffic facilities, catering facilities, commercial facilities, and park green space in the streets, and limit the distribution density of POI in a given area. Furthermore, in order to extend the time for pedestrians to carry out leisure walking behavior in the living streets of the old city districts, it is possible to increase the layout of park greening facilities and living service facilities of living streets in old urban areas and pay attention to the diversity of street greening and street landscape. In addition, in order to promote the city–environment–health degree of old urban areas in high-latitude cities, the multi-factor framework proposed in this paper can be used to quantify the potential attraction to pedestrians when replanning and designing new streets, so as to maximize the effectiveness and efficiency of decision making. These contributions contribute to improving the current situation of people-oriented living streets environment planning and guide landscape architects, urban designers and policy makers to establish healthy, safe and positive environments for pedestrian streets in old urban districts.

In fact, the living street is the type most closely related to the lives of its residents. There are many types of streets in a city, and the activities carried out by pedestrians on the streets are not just walking behavior. This study aims to study the influence of the living street environment on pedestrian walking behavior. In the future, it would be interesting to classify pedestrian walking behavior in more detail than just traffic and leisure walking. At the same time, a comprehensive classification of the different behavioral activities of pedestrians will be the focus of the next study on the impact of the living street environment in old urban districts on pedestrian behavior. We also hope that this study can be extended to other street types in future research, and that the classification of street environments will not be limited to visual perception environments. The acquisition of data is the factor that currently limits us to make detailed classification. The development of methodologies to classify pedestrian walking behavior using objective data will be a focal point of subsequent studies. The insights gleaned from such research will equip landscape architects, urban designers, and policy makers with the necessary knowledge to create health-conscious, secure, and positive environments for pedestrian streets within historic urban districts.