The Study on Spatial Elements of Health-Supportive Environment in Residential Streets Promoting Residents’ Walking Trips

Residents’ walking trips are a kind of natural motion that promotes health and wellbeing by modifying individual behavior. The purpose of this study was to evaluate the major influence of the spatial elements of a health-supportive environment on residents’ walking trips. This study analyzes residents’ walking trips’ elements based on the spatiotemporal characteristics of walking trips, as well as the spatial elements of a health-supportive environment in residential streets based on residential health needs. To obtain the spatial elements that promote residents’ walking trips and to build an ordered logistic regression model, two methods—a correlation analysis and a logistic regression analysis—were applied to analyze the elements of residents’ walking trips as well as the spatial elements of a health-supportive environment in residential streets by means of SPSS software, using on-site survey results of ten residential streets and 2738 pieces of research data. The research showed that the nine kinds of spatial elements that significantly affect residents’ walking trips are density of pedestrian access, density of bus routes, near-line rate of roadside buildings, average pedestrian access distance, square area within a 500 m walking distance, distance to the nearest garden, green shade ratio, density of street intersections, and the mixed proportion of differently aged residential buildings. In order to construct a spatial environment that promotes walking trips, it is necessary to improve the convenience of residents’ walking trips, to increase the safety of roadside buildings and pedestrian access, to expand the comfort of “getting out to the nature”, and to enrich the diversity of different architectural styles and street density.


Introduction
Physical activity (PA) is a kind of generally recognized protective factor. However, physical inactivity (PI) has become a universal problem around world [1]. According to statistics, 31.1% of the world's adults cannot reach the standard level of physical activity. The China National Health and Planning Commission (CNHPC) found that the regular exercise rate of adults is only 18.7% in China [2]. Many studies have shown that physical inactivity (PI) has a substantial influence on obesity, heart disease, hypertension and diabetes, among others. It has become the most serious challenge of global public health in the 21st century [3][4][5][6][7]. The question of how to scientifically and reasonably intervene and improve the physical activity of residents is particularly important for improving the physical health of the general population; it is also of particular importance for the "Healthy China" (HC) strategy [8,9]. Improving the health level of residents via environmental interventions is the key aim of Healthy City Construction (HCC) in China. Constructing a health-supportive environment is

Research Area
The Nan'an District, in Chongqing City, was selected as the research area for this study. The Nan'an District is the core area in Chongqing city, covering an area of 262.43 square kilometers, with a resident population of 530,000 people. It is surrounded by the Yangtze River to the north and west. The overall terrain is relatively flat, with a few gentle slopes, as is common in mountainous cities. To avoid an impact of the business district on residents' walking trips, ten residential streets within a straight-line distance of more than 1 km from Nanping business district were selected as the research samples ( Figure 1). The research area's scope is the street space within a 500 m walking path from the sampled residential area. At present, most of the residential areas in China are access-controlled residential areas with walls, and residents and visitors must be authenticated before entering the residential area. In Figure 1, the yellow area is the range of the residential area, the green dot is the pedestrian access of the residential area, and the red line represents the 500 m walking path starting from the pedestrian access of the residential area. Residents' walking trips' data from the sample residential areas were gathered using on-site surveys and questionnaires. Therefore, there was a large amount of data on residents' walking trips in one residential area. Spatial elements' data from the sampled residential areas were gathered using an on-site survey, street view maps, and building mapping. Therefore, for each sample residential area there was one dataset regarding spatial elements. In order to solve the problem that one spatial element dataset corresponds to data pertaining to multiple residents' walking trips in each sampled residential area, the "mode" of multiple residents' walking trips data in the sample residential area was taken as residents' walking trips data of the sample residential area so as to achieve a one-to-one correspondence between the residents' walking trips data and the spatial elements' data in the research (Figure 2). Residents' walking trips' data from the sample residential areas were gathered using on-site surveys and questionnaires. Therefore, there was a large amount of data on residents' walking trips in one residential area. Spatial elements' data from the sampled residential areas were gathered using an on-site survey, street view maps, and building mapping. Therefore, for each sample residential area there was one dataset regarding spatial elements. In order to solve the problem that one spatial element dataset corresponds to data pertaining to multiple residents' walking trips in each sampled residential area, the "mode" of multiple residents' walking trips data in the sample residential area was taken as residents' walking trips data of the sample residential area so as to achieve a one-to-one correspondence between the residents' walking trips data and the spatial elements' data in the research ( Figure 2).

Study Characteristics and Study Methods
Residents' walking trips have spatiotemporal characteristics. The time characteristics are the major evaluation criteria of residential health. The spatial characteristics provide the space for health needs. In this study, residents' walking trips' elements were analyzed on the basis of their spatiotemporal characteristics, and the spatial elements of a health-supportive environment in residential streets were analyzed based on residential health needs. To obtain the spatial elements that promote residents' walking trips, two methods, including correlation analyses and logistic regression, were applied to analyze walking trips elements and spatial elements of health-supportive environments in residential streets ( Figure 3). In 2018, research data were obtained using an on-site survey and a questionnaire survey (Appendix A, Table A1) analyzing residents' walking trips. The interview and questionnaires were administered to residents who were entering and exiting the residential area at the pedestrian entrances and exits of the access-controlled residential area to ensure that the survey objects were residents of the same residential area. The numbers of questionnaires for each sample residential area were proposed to be 10% of the total numbers of households in the sample residential area, with 2738 questionnaires (Table 1). 2231 valid questionnaires were collected and the statistical data were as shown in Tables 2 and 3:

Study Characteristics and Study Methods
Residents' walking trips have spatiotemporal characteristics. The time characteristics are the major evaluation criteria of residential health. The spatial characteristics provide the space for health needs. In this study, residents' walking trips' elements were analyzed on the basis of their spatiotemporal characteristics, and the spatial elements of a health-supportive environment in residential streets were analyzed based on residential health needs. To obtain the spatial elements that promote residents' walking trips, two methods, including correlation analyses and logistic regression, were applied to analyze walking trips elements and spatial elements of health-supportive environments in residential streets ( Figure 3).  One-to-one data correspondence.

Study Characteristics and Study Methods
Residents' walking trips have spatiotemporal characteristics. The time characteristics are the major evaluation criteria of residential health. The spatial characteristics provide the space for health needs. In this study, residents' walking trips' elements were analyzed on the basis of their spatiotemporal characteristics, and the spatial elements of a health-supportive environment in residential streets were analyzed based on residential health needs. To obtain the spatial elements that promote residents' walking trips, two methods, including correlation analyses and logistic regression, were applied to analyze walking trips elements and spatial elements of health-supportive environments in residential streets ( Figure 3). In 2018, research data were obtained using an on-site survey and a questionnaire survey (Appendix A, Table A1) analyzing residents' walking trips. The interview and questionnaires were administered to residents who were entering and exiting the residential area at the pedestrian entrances and exits of the access-controlled residential area to ensure that the survey objects were residents of the same residential area. The numbers of questionnaires for each sample residential area were proposed to be 10% of the total numbers of households in the sample residential area, with 2738 questionnaires (Table 1). 2231 valid questionnaires were collected and the statistical data were as shown in Tables 2 and 3: In 2018, research data were obtained using an on-site survey and a questionnaire survey (Appendix A, Table A1) analyzing residents' walking trips. The interview and questionnaires were administered to residents who were entering and exiting the residential area at the pedestrian entrances and exits of the access-controlled residential area to ensure that the survey objects were residents of the same residential area. The numbers of questionnaires for each sample residential area were proposed to be 10% of the total numbers of households in the sample residential area, with 2738 questionnaires (Table 1). 2231 valid questionnaires were collected and the statistical data were as shown in Tables 2 and 3:   People who chose "moderate exercise" were administered a further in-depth interview. A total 375 of people participated in the in-depth interview, 83.1% of the which agreed with and practiced the idea of residents' walking trips to promote health; they were further investigated regarding their health needs, the spatial environment of residential streets, and the current situation of residents' walking trips.

Analysis of Elements of Residents' Walking Trips
Spatiotemporal characteristics are the main research perspective regarding residents' walking trips. According to the statistics of the questionnaire, the age of the survey objects was mostly concentrated in the middle-aged and young groups (62.1%), of which fixed workers accounted for a relatively high proportion (60.9%).
The time required for getting to work (one way) was within 10 min (8.3%), 10-20 min (14.1%), 20-30 min (27.9%), 30-40 min (23.50%), 40-50 min (13.2%), 50-60 min (9.1%), and more than 60 min (3.9%). Most residents' one-way commute time to work was within 20-40 min; the smallest group had a one-way commute to work that took more than 60 min. Therefore, taking the daily walking and public transportation commuters as an example, the time-space prism simulation can be more representative for the analysis of elements of residents' walking trips (The spatiotemporal prism methodis a method of modeling the physiological, physical, and environmental constraints on individual behavior [31]).
Necessary activities refer to activities restricted by activity location and activity time; triggered activities refer to the activities that can be randomly selected within limits. There are many triggered activities that may arise in the process of performing necessary activities. As shown in Figure 4, there are multiple triggered activity locations (such as breakfast shops, stores, supermarkets and other triggered activity locations) in the necessary activities locations (such as schools, homes, workplaces and other necessary activity locations). Residents' walking trips can be divided into necessary residents' walking trips with a specific destination and with limited time, and triggered walking trips without a clear destination and with more spatial possibilities.   Residents' daily walking trips' time is positively correlated with their health situation [32], which can be used as a scale to evaluate the health-supportive environment in residential streets. Through the questionnaire survey from the ten sampled residential areas, it was found that walking time and the frequency of walking trips were not only limited by the street space, but also affected by residents' daily schedules. The total individual walking trips' time within a certain period was determined by the frequency of walking trips and their durations together.
The frequency of walking trips includes the frequency of necessary walking trips and the frequency of triggered walking trips. The duration of walking trips includes the duration of necessary walking trips and the duration of triggered walking trips. Therefore, the elements of residents' walking trips mainly include four kinds as follows: the frequency of necessary walking trips defined as Y1, the duration of necessary walking trips defined as Y2, the frequency of triggered walking trips defined as Y3, and the duration of triggered walking trips defined as Y4 (Table 4). One to two times per week-value "1" Three to six times per week-value "2" One time per day-value "3" Two times per day and above-value "4"

Triggered activities
Activities that can be randomly selected within limits Frequency of residents' triggered walking trips (Y3) One to two times per week-value "1" Three to six times per week-value "2" One time per day-value "3" Two times per day and above-value "4"

Residential Health Needs
Based on residential health needs, this study discusses the elements of residents' walking trips in residential streets; a health-supportive environment refers to an environment that can meet physical health needs, mental health needs, social adaptability health needs, and moral health needs [33]. Therefore, the residential health needs can be divided into physical health needs, mental health needs, social adaptability health needs, and moral health needs from the perspectives of residents' walking trips as follows: Physical health needs mainly refer to a spatial environment that is convenient for walking trips in residential streets; this includes smooth walking paths, good pedestrian access, and bus stations, and an environment that is more convenient for business facilities [32,33].
Mental health needs mainly refer to a spatial environment in residential streets that is conductive to the release of residents' psychological pressure; a higher density of frontage stores; an appropriate height-to-width ratio of the street; wide pavements; and beautiful street greening [34].
Social adaptability health needs mainly refer to a spatial environment in residential streets that promotes good neighborhood relationships, including higher density of street intersections, multi-function streets, mixed building ages, multi-functional shops near the street, and high walkway flow [35].
Moral health needs mainly refer to a spatial environment in residential streets that protects pedestrians. Pedestrians are a vulnerable group compared to motor vehicles, and there are also relatively vulnerable groups among pedestrians, such as the sick and the disabled. In order to protect pedestrians from the interference of motor vehicles, enough pedestrian crossing facilities, enough outdoor facilities, and sufficient night lighting are needed. In order to protect the walking convenience of the disabled, a perfect blind track system and wheelchair-accessible ramps are needed [36].

The Current Problems in Residential Streets
Based on the on-site interviews and questionnaires, it was found that the spatial environment in the sampled residential streets cannot completely meet the residential walking health needs. From the interviews regarding the two aspects of "the spatial environment promoting residents' walking trips in residential streets" and "the spatial environment hindering residents' walking trips in residential streets" (Appendix A, Table A2), four main problems at present were found, as follows: Path problems of walking trips: low density of the residential street network; unsmooth pedestrian paths; narrow width of pedestrian walkways on streets; and no openings or few openings on the street interface of the building.
Safety problems of walking trips: slippery road surfaces in rainy days; inadequate lighting at night; lack of crossing facilities; illegal occupation of sidewalks.
Environmental problems of walking trips: lack of humanization of hard pavement; poor aesthetic effect of buildings along the street; lack of green landscape; imbalance of height-width ratio of street space; poor environmental sanitation.
Spatial function problems of walking trip streets: lack of shared leisure communication spaces and rest nodes; lack of effective connections between the underlying space of the building and the street; fragmentation of the street space; lack of open shops on the street; lack of places suitable to activities for differently aged.

Spatial Elements of a Health-Supportive Environment in Residential Streets Based on Residential Health Needs
The above four main problems correspond to the health needs of residents' walking trips, which is an important correspondence for researching the health-supportive environment in residential streets. The spatial elements of a health-supportive environment in residential streets can be divided into convenience, safety, comfort, and diversity ( Figure 5).
Safety: traffic safety [44,45], which includes the location and length of crosswalks, traffic signs and signals for pedestrians and vehicles, and width of sidewalks [37,46,47]; public security, which includes road monitoring, night lighting, social security, and personal safety; psychological safety [37,48], which includes street closures, facade transparency of ground floor buildings, the proportion of ground floor commercial buildings, and the density of stores along the street [47,49].
Diversity: the distance between leisure and entertainment shopping destinations [41,42] which includes the density of commercial facilities and the density of leisure and entertainment facilities [51,57,58]; various land-use models, which includes the mixed use of frontage stores and the varieties of street forms [45,59,60].
Int. J. Environ. Res. Public Health 2020, 17, x 10 of 33 residential streets. The spatial elements of a health-supportive environment in residential streets can be divided into convenience, safety, comfort, and diversity ( Figure 5). Convenience: connectivity of residents' walking trips network, which includes intersection density and public transport network density [37][38][39]; and accessibility of public facilities [37,40], including a comprehensive supermarket coverage radius and a leisure and entertainment coverage radius [41][42][43].
Safety: traffic safety [44,45], which includes the location and length of crosswalks, traffic signs and signals for pedestrians and vehicles, and width of sidewalks [37,46,47]; public security, which includes road monitoring, night lighting, social security, and personal safety; psychological safety [37,48], which includes street closures, facade transparency of ground floor buildings, the proportion of ground floor commercial buildings, and the density of stores along the street [47,49].
Diversity: the distance between leisure and entertainment shopping destinations [41,42] which includes the density of commercial facilities and the density of leisure and entertainment facilities [51,57,58]; various land-use models, which includes the mixed use of frontage stores and the varieties of street forms [45,59,60].

The Spatial Element Data Collection of A Health-Supportive Environment in Residential Streets
On the basis of research relevant to this study, the spatial element data were collected through on-site interviews, street view maps, and building surveying and mapping. The specific process was as follows: First of all, the relevant data of spatial elements were obtained from the CAD file of the current land use situation of Nan'an District, Chongqing City, and coordinated with streetscape maps, Baidu maps and Tencent maps to obtain relevant data of spatial elements (Appendix A, Table A3).
Secondly, through on-site interviews, the obtained data were supplemented, perfected, and amended to improve the accuracy and the reliability of the original data.
Finally, the spatial elements of the health-supportive environment in residential streets were subdivided, numbered, and assigned in the study (Appendix A, Table A4).

The Spatial Element Data Collection of a Health-Supportive Environment in Residential Streets
On the basis of research relevant to this study, the spatial element data were collected through on-site interviews, street view maps, and building surveying and mapping. The specific process was as follows: First of all, the relevant data of spatial elements were obtained from the CAD file of the current land use situation of Nan'an District, Chongqing City, and coordinated with streetscape maps, Baidu maps and Tencent maps to obtain relevant data of spatial elements (Appendix A, Table A3).
Secondly, through on-site interviews, the obtained data were supplemented, perfected, and amended to improve the accuracy and the reliability of the original data.
Finally, the spatial elements of the health-supportive environment in residential streets were subdivided, numbered, and assigned in the study (Appendix A, Table A4).

The Quantitative Analysis Method of Spatial Elements for a Health-Supportive Environment that Promotes Walking Trips in Residential Streets
The spatial elements of a health-supportive environment are defined as X variables (Appendix A, Table A4), and the elements of residents' walking trips are defined as Y variables (Appendix A, Table A4). Data processing and data modeling were carried out for X and Y variables. The quantitative analysis of spatial elements was divided into two steps as follows ( Figure 6).

The Quantitative Analysis Method of Spatial Elements for A Health-Supportive Environment that Promotes Walking Trips in Residential Streets
The spatial elements of a health-supportive environment are defined as X variables (Appendix A, Table A4), and the elements of residents' walking trips are defined as Y variables (Appendix A, Table A4). Data processing and data modeling were carried out for X and Y variables. The quantitative analysis of spatial elements was divided into two steps as follows ( Figure 6). Step 1: Correlation analysis. To explore the spatial elements of a health-supportive environment in residential streets that promotes residents' walking trips, we had to find a X variable that had a significant correlation with the Y variable, using the methods of a contingency coefficient analysis, a Spearman's coefficient analysis, and a second choice of variable clustering. Different variable types corresponded to different correlation analysis methods and judgment standards (Table 5). ; ordinal variable: the data are a group of numbers, and the grades between the numbers increase or decrease, such as the satisfaction variable ("1" dissatisfaction, "2" neutral, "3" satisfaction); category variable: the data are a group of numbers, and these numbers are not related, such as the gender variable ("1" male, "2" female).
For the X and Y variables filtered in step 1, the X variable is defined as an independent variable and the Y variable is defined as a dependent variable. Quantitative analysis and model fitting equation calculations can be carried out using ordered logistic regression analysis. The model and specific hypothesis of the ordered logistic regression are cumulative ordered logistic models, which need to be calculated using a cumulative probability formula (Formulas 1-4). For some X variables Step 1: Correlation analysis. To explore the spatial elements of a health-supportive environment in residential streets that promotes residents' walking trips, we had to find a X variable that had a significant correlation with the Y variable, using the methods of a contingency coefficient analysis, a Spearman's coefficient analysis, and a second choice of variable clustering. Different variable types corresponded to different correlation analysis methods and judgment standards (Table 5). ; ordinal variable: the data are a group of numbers, and the grades between the numbers increase or decrease, such as the satisfaction variable ("1" dissatisfaction, "2" neutral, "3" satisfaction); category variable: the data are a group of numbers, and these numbers are not related, such as the gender variable ("1" male, "2" female).
For the X and Y variables filtered in step 1, the X variable is defined as an independent variable and the Y variable is defined as a dependent variable. Quantitative analysis and model fitting equation calculations can be carried out using ordered logistic regression analysis. The model and specific hypothesis of the ordered logistic regression are cumulative ordered logistic models, which need to be calculated using a cumulative probability formula (Formulas (1)-(4)). For some X variables that cannot be calculated by the formula calculation, an OR value can be used to judge their influence on the Y variables (Formula (5)).
In the cumulative probability formula, the dependent variable j was set as the ordered variable of K pieces of grades, and K pieces of grades was represented by 1, 2 . . . K. Grade K is divided into two categories: {1, J} and {j + 1, K}, while j is the cutting point; the logit is defined on the basis of these two types: the logarithm of the advantages of the accumulated probability of the last K-J level and the accumulated probability of the first j level; the model is thus called the accumulated advantage model, in which the model assumes that the level of the dependent variable is the same regardless of classification. That is, the regression coefficient of the independent variable is independent of the partition point J, which is the cumulative probability ratio of the former to the latter: Logit P j = Logit P(y > j x) = ln P(y > j x) The OR value indicates the ratio of the y value to the ratio of one or more grades (the logarithm of the dominance ratio) for each unit of change of independent variable Xi; that is, how many times does the y value increase for each unit increase of the independent variable Xi.):

Results
In the process of the statistical data analysis, the data were analyzed from the correlation analysis and the ordered logistic regression in this study. The research results include the results of the correlation analysis and the results of the ordered logistic regression analysis.

Correlation Analysis Results
The significance level test was set to 0.05, and the correlation analysis conclusions were as follows: The spatial elements needed to improve the frequency of necessary residents' walking trips (Y1) were as follows ( Table 6): density of pedestrian access (XB1), convenience of service stores (XB4), density of bus routes (XB11), and average pedestrian access distance (XC3).
The spatial elements needed to increase the duration of necessary residents' walking trips (Y2) were as follows ( Table 6): mixed uses of service store kinds (XA3), density of street intersections (XA15), density of street directions (XA16), mixed proportion of differently aged of residential buildings (XA17), ratios of old residential buildings (XA19), and average pedestrian access distance (XC3).
The spatial elements needed to improve the frequency of triggered residents' walking trips (Y3) were as follows ( Table 6): mixed uses of leisure and entertainment store kinds (XA4), population density (XA20), density of pedestrian access (XB1), convenience of leisure and entertainment store kinds (XB10), average pedestrian access distance (XC3), square area within a 500 m walking distance (XD1), and distance to the nearest garden (XD2).
The spatial elements needed to increase the duration of triggered residents' walking trips (Y4) were as follows ( Table 6): mixed uses of service store kinds (XA3), distribution density of leisure and entertainment stores (XA10), density of street intersection (XA15), density of street directions (XA16), ratios of old residential buildings (XA19), density of pedestrian access (XB1), convenience of service stores (XB4), convenience of service store kinds (XB9), near-line rate of roadside buildings (XC2), average pedestrian access distance (XC3), and green shade ratio (XD6).
In summary, the conclusions are as follows: safety is the basis of walking trips; this has an impact on the frequency of necessary walking trips, the duration of necessary walking trips, the frequency of triggered walking trips and the duration of triggered walking trips. Meanwhile, convenience has the greatest influence on the frequency of necessary and triggered walking trips. Diversity has the greatest influence on the duration of necessary and triggered walking trips. Comfort has the greatest influence on the frequency and duration of triggered walking trips.

Ordered Logistic Regression Analysis Results
The significance level test was set to 0.05, the Y variable was defined as the dependent variable and the X variable was defined as the independent variable. The ordered logistic regression analysis conclusions are as follows: The spatial elements needed to improve the frequency of necessary residents' walking trips (Y1) are as follows ( Table 7): density of pedestrian access (XB1), density of bus routes (XB11), and average pedestrian access distance (XC3).
The spatial elements needed to increase the duration of necessary residents' walking trips (Y2) are as follows ( Table 7): density of street intersection (XA15), and mixed proportion of differently aged residential buildings (XA17).
The spatial elements needed to improve the frequency of triggered residents' walking trips (Y3) are as follows (Table 7): square area within a 500 m walking distance (XD1) and distance to the nearest garden (XD2).
The spatial elements needed to increase the duration of triggered residents' walking trips (Y4) are as follows (Table 7): near-line rate of roadside buildings (XC2), average pedestrian access distance (XC3), green shade ratio (XD6).

Insights from the Correlation Analysis regarding the Construction of a Health-Supportive Environment in Residential Streets
There are many relatively comprehensive studies on the spatial elements of residential streets that promote residents' walking trips: the European alpha environmental questionnaire from 2004 and the US residents' walking trips index from 2007 and 2013 [30]. However, the impact of spatial elements on residents' walking trips has not yet been fully investigated. Based on the results of the correlation analysis, this study effectively quantifies the impact degree of each spatial element on walking trips and analyzes the impact degree by the frequency level (Figure 7). Thirdly, public open space and the types of leisure and entertainment also have a certain impact on walking trips. Increasing the distribution density and distribution species of leisure and entertainments facilities should be considered, as well as improving the greening coverage and increasing the density of parks and the squares during urban planning and construction.
Finally, the density of bus routes and population density also have a certain impact on walking trips. They are part of the contents of urban master planning. The density of bus lines should be increased and the population density should be improved in urban master planning in the future.

Application of the Ordered Logistic Regression Analysis to A Health-Supportive Environment in Residential Streets
First, a health-supportive environment evaluation standard on the basis of walking trip time as a health standard was built; second, ten sample residential areas were evaluated by using walking trips time as a health standard; and finally, the effective adjustment of the spatial elements used the ordered logistic regression formula.

Establishing A Health Evaluation Standard
Residents' physical health was measured by the number of walking steps or daily walking time [61,62]; the walking health standard was set in this study as 60 min or more of walking every day [63][64][65][66].

Health Evaluation of the Ten Sampled Residential Areas
Seven sampled residential areas were unqualified during evaluating in ten sampled residential areas (Table 8).

The Effective Adjustment of the Spatial Elements Used in the Ordered Logistic Regression Formula
The spatial elements of seven unqualified sample residential areas were adjusted effectively by using the ordered logistic regression formula to increase the residents' walking trip times per day.
Jinyang Roman holiday residential street was taken as an example: • Model test ( Table 9): The predicted data of the model were consistent with the real data, and the model can be further analyzed. There are some insights from the correlation analysis regarding health-supportive environmental construction in residential streets: First, pedestrian access has a great impact on the residents' walking trips in residential streets. Non-gated residential communities and small-scale residential communities can effectively promote walking trips. Therefore, more pedestrian accesses will be considered in urban planning and construction.
Second, roadside buildings, street intersections, service facilities, and so on have great influence on walking trips. The density of roadside buildings should be increased, T-shaped road breaking roads should be avoided, the openness of roadside buildings should be improved, buildings with historical characteristics should be maintained, the functions of roadside buildings should be enriched, and the distribution density and the distribution species of service facilities should be increased in the process of urban planning and construction.
Thirdly, public open space and the types of leisure and entertainment also have a certain impact on walking trips. Increasing the distribution density and distribution species of leisure and entertainments facilities should be considered, as well as improving the greening coverage and increasing the density of parks and the squares during urban planning and construction.
Finally, the density of bus routes and population density also have a certain impact on walking trips. They are part of the contents of urban master planning. The density of bus lines should be increased and the population density should be improved in urban master planning in the future.

Application of the Ordered Logistic Regression Analysis to a Health-Supportive Environment in Residential Streets
First, a health-supportive environment evaluation standard on the basis of walking trip time as a health standard was built; second, ten sample residential areas were evaluated by using walking trips time as a health standard; and finally, the effective adjustment of the spatial elements used the ordered logistic regression formula.

Establishing a Health Evaluation Standard
Residents' physical health was measured by the number of walking steps or daily walking time [61,62]; the walking health standard was set in this study as 60 min or more of walking every day [63][64][65][66].

Health Evaluation of the Ten Sampled Residential Areas
Seven sampled residential areas were unqualified during evaluating in ten sampled residential areas (Table 8).

The Effective Adjustment of the Spatial Elements Used in the Ordered Logistic Regression Formula
The spatial elements of seven unqualified sample residential areas were adjusted effectively by using the ordered logistic regression formula to increase the residents' walking trip times per day.
Jinyang Roman holiday residential street was taken as an example: • Model test ( Table 9): The predicted data of the model were consistent with the real data, and the model can be further analyzed.
When the frequency level of residents' necessary walking trips (Y1) and the frequency level of residents' triggered walking trips (Y3) were adjusted effectively, residents' walking trip times per day were adjusted from 17 to 60 min.
• Spatial element index adjustment (Table 11): The density of pedestrian access, the density of bus routes, the average pedestrian access distance, the distance to the nearest block parks and the area of squares accessibility within a 500 m distance were adjusted with the ordered logistic regression Formulas (7)-(10) (Section 3.2.1) and Formulas (18)-(21) (Section 3.2.3). However, this adjustment is only for theoretical guidance and is combined with the current situation in the practice application.

Research Limitations
There are certain errors in the statistical data of residents' walking trips, which were gathered by means of on-site interviews and questionnaires. Therefore, for further research, large data and mobile devices are considered for data sample collection of residents' walking trips.
The residents who enter and exit the residential area at the pedestrian access points are mobile; this survey did not conduct more in-depth household surveys. Therefore, it is true that more than one person in the same household may have participated in the research, but it is the small probability event.
There is no direct related data analysis between the spatial elements of a health-supportive environment and residents' health status.
In the process of the quantification of abstract data and the statistical operation of abstract data, it is inevitable to simplify and neglect the personality of residents, such as residents' age and gender, which led to some deviations in the results.
The reconstruction plan proposed by the regression model is only for theoretical guidance, and there may be some obstacles in practice, such as the increase of pedestrian access and bus routes, which need the coordination of the government, community, and residents.   Forecast data Y1 = "2" Y2 = "4" Y3 = "1" Y4 = "4" Real data Y1 = "2" Y2 = "4" Y3 = "1" Y4 = "4"  1 The data of the frequency of residents' walking trips and the duration of necessary residents' walking trips are calculated and adjusted according to the minimum value. Table 11. Adjustment of spatial elements.

Adjustments of Spatial Elements (X) Corresponding to the Frequency of Necessary Residents' Walking Trips (Y1) Original Data Modified Data
Independent variable (X)

Specific Research for Particular Groups
The walking trip characteristics of residents of different genders or different age groups are different. The relationship between residents' walking trips and residential street space environment can be studied according to different genders or different age groups in the future.

Research in Depth on a Sample Residential Area
The transformation strategy guided by theory is applied to the practice of residential street environment construction. Through a more in-depth household survey and the feedback of the project, the optimization and modification of the health-supportive environment for the health of the residential streets to promote walking trips are continuously strengthened.

Follow-Up Sample Population in the Long Term
After using the ordered logistic regression model to adjust the spatial elements of the sample residential area, the impact of the adjusted spatial elements on the walking trip times and the health of sample residential area residents can be explored.

Conclusions
The spatial elements of a health-supportive environment in residential streets that promote residents' walking trips and the transformation strategy with the support of a data model for the improvement of residents' walking trips' environment in residential streets were studied in this paper mainly from three angles, as follows: Based on the spatiotemporal characteristics of residents' walking trips, the residents' walking trips' factors were divided into necessary walking trips' frequency, necessary walking trips' time, triggering walking trips' frequency, and triggering walking trips' time.
Based on the residents' health needs, the spatial elements of the health-supportive environment in residents' streets were analyzed and the spatial elements were refined from four angles: convenience, safety, comfort, and diversity.
Based on a correlation analysis and an ordered logistic regression analysis, this paper studied the spatial elements of a health-supportive environment for promoting residents' walking trips in residential streets, established an ordered logistic regression model, and improved residents' daily walking trips' time by adjusting spatial elements in reverse.
The study found that the main actions needed are as follows: improving the density of pedestrian access and the density of bus routes in residential areas to improve the convenience of residential streets; improving the near-line rate of roadside buildings and the average pedestrian access distance to increase the safety of residential streets; increasing the square area within a 500 m walking distance, the distance to the nearest garden, and the green shade ratio to improve the comfort of residential streets; increasing the density of street intersections and the mixed proportion of differently aged residential buildings to improve the diversity of residential streets.     Table A3. List of spatial element data collected with each method.

Method Spatial Element Data
CAD file of the current land use situation of Nan'an District, Chongqing City Area of the sample residential area (Hm 2 ) Walking distance within a 500 m walking path starting from the pedestrian access (km). Length of store-fronted streets (km) Square area within a 500 m walking path starting from the pedestrian access (Hm 2 ) Distance to the nearest garden away from the pedestrian access (   , where p i is the ratio of the use of No. i street-front stores within the street space, and n is the total numbers of uses of street-front stores

XA2
Mixed uses of shopping store kinds Mixed uses of shopping store kinds = − n i=1 pilnpi ln n , where p i is the ratio to the shopping uses of No. i kinds of street-front stores, and n is the total shopping kinds of street-front stores

XA3
Mixed uses of service store kinds Mixed uses of service store kinds = − n i=1 pilnpi ln n , where p i is the ratio to the service uses of No. i kinds of street-front stores, and n is the total service kinds of street-front stores

XA4
Mixed uses of leisure and entertainment store kinds Mixed uses of leisure and entertainment store kinds = − n i=1 pilnpi ln n , where p i is the ratio to the leisure and entertainment uses of No. i kinds of street-front stores, and n is the total leisure and entertainment types of street-front stores

XA5
Distribution density of catering stores Quantity of catering stores divided by walking distance within a 500 m walking path starting from the pedestrian access (Pcs/km)

XA6
Distribution density of shopping stores Quantity of shopping stores divided by walking distance within a 500 m walking path starting from the pedestrian access (Pcs/km)

XA7
Distribution density of shopping stores kinds Quantity of shopping stores kinds divided by walking distance within a 500 m walking path starting from the pedestrian access (Pcs/km)

XA8
Distribution density of service stores Quantity of service stores divided by walking distance within a 500 m walking path starting from the pedestrian access (Pcs/km)

XA9
Distribution density of service store kinds Quantity of service stores kinds divided by walking distance within a 500 m walking path starting from the pedestrian access (Pcs/km) , where p i is the ratio of residential buildings in all buildings, and n is the total types of residential building.

XA18
Whether or not there are old residential buildings 0 means it does not exist 1 means it exists

XA19
The ratio of old residential buildings Quantity of old residential buildings divided by quantity of all buildings XA20 Population density Total number of households divided by walking distance within a 500 m walking path starting from the pedestrian access (Pcs/km)