Kun Liu

Yuan Gao

Abstract

Physical activity (PA) is an essential way to promote the public health and has been considered as a vital indicator of urban health. How urban public space can provide equal opportunities for different groups of people and can support them to do various activities, is one of the core issues in urban planning. This paper focuses on PA in public spaces in Shenzhen, to analyze the diversities of PA types, PA time, and PA participants. This paper first creates two built environment indicators: general built environmental index, and particular built-environmental index for PA types, PA time, and PA participants, and then use the regression model to explore the association between the three types of diversity and the built environmental factors. The results show that the design of public space facilities and the inclusiveness of public space are the particular factor that significantly affect the diversity of PA type within the space. The all-day services and security level of public space are the core elements that can affect the diversity of PA time. The resource allocation of public space and the diversity of the surrounding housing types are core factors that can influence the diversity of PA participants’ class. In order to support a variety of activities, good accessibility and sizeable green space are essential conditions. Also, comprehensive sports facilities and venues, sizable shade and plenty of supportive facilities are also necessary.

Introduction

In recent years, a large number of urban renewal and development projects have actively advocated enhancing urban vitality, which has been proved to be an essential means to optimize urban quality and promote the development of cities (Zhang, Zhang & Zhou 2017). Jane Jacobs pointed out that it is necessary to advocate the development of "public space" in "urban space", and regard public space as the key factor to promote the formation of good social interaction and the restoration of urban vitality in urban development and community construction. (Jacobs 1961) In this context, the land use mode with diverse uses and functions has replaced the traditional single functional area. The planning and construction of urban public space also begin to focus on the demands of multi-crowd and multi-time dimensions, as the vitality of the city largely depends on whether the city can satisfy people’s diverse demands.

1.1 Literature review

Previous studies have demonstrated that a well-designed built environment has a positive effect on promoting people’s PA (Weng, He, Wang, Lin & Li, 2010; Zhou, Li & Fu, 2012). The core factor of the built environment which affect the PA are the design of street, land use, the type of community, and aesthetics. (Zapata Diomedi & Veerman, 2016; Wang & Lee, 2010; Mass & Verheij, 2007; Lin & Yang, 2015). As an important part of the built environment, residential density is one of the critical environmental attributes related to PA (Forsyth, Oakes & Schmitz,2007；Frank 2000；Cervero & Kockelman,1997). A large number of studies have shown that the characteristics of residential areas can influence PA. For example, people who live close to workplace, service place and shop will have a higher probability of doing walking and cycling activities(Zapata Diomedi & Veerman, 2016).

The researches about the relationship between the built environment and diversity of PA mainly focus on the diversity of PA type and the diversity of PA participants’ class. The types of PA include walking, jogging, and cycling. Moreover, the land use mixture, safety lexel, environmental aesthetics, accessibility, distribution of facility, street connectivity, and the ability to provide recreational opportunities are the influencing factors for people to do walking and cycling activities. Beautiful scenery can attract people to do PAs such as walking and cycling (Wang, Chau & Ng, 2016; Inoue, Ohya,&Odagiri, 2010; Inoue, Murase & Shimomitsu, 2009). Street network characteristics have a significant impact on walking, cycling and jogging activities of adults (Orleans, Kraft & Marx, 2003). Walkability is positively correlated with the type and duration of weekly PA (Rundle, Sheehan & Quinn, 2015). And in low urbanization areas, street density is positively associated with the frequency of three types of activities (Hou, Popkin & Jacobs, 2010). The green place that with a mixed layout of multiple functions, a convenient public transport environment, and a strong continuity can attract more diverse PAs（Liu, Tan & Song, 2016; Feng & Yang, 2015）.

The built environment affects the subjective perceptions of people of different ages, genders, social and economic status, and so on, and thus influence the level of PA（Wang & Lee, 2010; Wang & Han, 2012）. Residential density and walkability index were correlated with middle-intensity and high-intensity PA among adolescents, and students (ages 13-15) living in low-residential density communities were more likely to engage in middle-intensity and high-intensity PA than students living in higher-density residential communities（Xu, Li & Liang, 2010）. High land use mixture and compact development mode can encourage older people to do PAs（Feng & Yang, 2015）. Some studies have shown that the level of PA tends to increase with the increase of educational level and income（Lee, Cubbin & Winkleby, 2007）, and housing costs are often positively related to the surrounding infrastructure, the coverage of park, safety level, and community cohesion（Jiang, Zhou & Xiao, 2010）. The public space around the residents is the priority choice for different classes to do activities（Zou, 2015）. People of different social classes participate PA in different time of days. In addition, the street scale has an impact on walking activities of different income and ethnic groups（Kelly, Schootman & Baker, 2007; Heinrich, Lee & Suminski, 2007）.

Most of the traditional researches of the built environment and PA usually concentrate on only one kind of activity（Karusisi, Bean & Oppert,2012；Boer, Zheng & Overton,2007； Zhang, Chen & Liu, 2010）. Nevertheless, a well-designed built environment should support various types of PAs（Liu, Siu, Gong, 2016）. In-depth study about the various demand of the PA is therefore, indispensable for the construction of the urban vitality, which will to promote the vitality of the city.

1.2 Diversity

Diversity is a philosophical concept which means that material elements form a complex and changeable material world through different forms. The essence of diversity in cities is to satisfy the demands of different groups and to form a balanced and harmonious spatial order（Pan, Jin & Liu, 2007）. Therefore, the diversity of PA in cities also includes different aspects. This paper mainly discusses the diversity of PA in three aspects: the diversity of PA type, the diversity of PA time, and the diversity of participants’ class.

2. Methods

2.1 Study area

This study chooses Shenzhen as the research area. Shenzhen is located in the eastern coast of the Pearl River Delta, in the south of Guangdong province, with an area of 1,991.64 square kilometers, and the annual average temperature is about 72.32℉, which is suitable for outdoor activities all around the year. The traffic network, public transportation system can cover almost all areas of Shenzhen; green space in the city is roughly 1214.95 square kilometers; the natural resources are relatively abundant (Fig 1).

People often do PAs in public space, such as roads, parks, residential areas, and other open spaces. A lot of studies use the method of grid partition to present and analyze the spatial data（Zhou, Wang & Ma, 2008, Li, Du & Weng, 2008）. Since the space of PAs is not unique, in order to accurately measure the diverse characteristics and the impact of the built environment on it, this paper will divide the space grid with 500 meters as the grid unit size. The attributes of diversity were explored through the grid, and then explore the impact of the built environment on the PA.

Fig 1 Aerial photograph of Shenzhen

2.2 Data resources

2.2.1 Physical activity data

In this study, the PA data were collected from Codoon, one of the most popular self-tracking smartphone applications in China. With Codoon, people can track their travel routes and record their activity date, type (classified as walking, jogging, and cycling), distance, speed, and duration. Also, the users can share the data on the website of Codoon. Based on the previous analyses of the data, the frequency of physical activity was at a peak in April and July. Therefore, the PA data in April and July 2015 in Shenzhen was selected for this study. Moreover, this paper uses the housing cost to indicate the social class of participant, to accurate the information of social class, we select routes starting or ending in residential areas and use the house price of the area to indicate their income level, which left 735 unique cases, walking, jogging, and cycling, was 194, 486, and 55 respectively.

2.2.2 Built environment data

In consideration of the impact of the built environment on PA, built environment data in this study contains land use data, street network data, greenway network data, the geographic information of the bus station data (GIB), the Normalized Vegetation Index (NDVI), and The point of interest data (POI).

Tab 1 Built environment data

2.2.3 Shenzhen housing data

This paper uses the housing price and rent price to inferring the social class of the participant of the physical activity. The housing price and the rent price of Shenzhen are from the housing information of Shenzhen (2016), which was published by Lianjia.com (China's leading real estate service platform). By excluding the incomplete information, 3647 housing price data and 1571 rent price data were selected to do the research. At the same time, the information of the city village, low-rent housing, public rental housing, and all kinds of affordable housing is also considered.

All the data are geocoded, calibrated and integrated into ArcGIS10 according to Shenzhen local coordinate system, forming the database of this study.

2.3 Measures

2.3.1 Independent variable

2.3.1.1 Descriptive analysis of physical activity data

Considering the activities in workday and weekday are different, and the temperature will influence the activities, this study collected the activity data of two working days and two rest days in April and July of 2015 from Codoon. Finally, 735 piece of activity data were selected, including 194 piece of walking data, 486 piece of jogging data and 55 piece of cycling data. All these data with starting or ending points as residential land. The essential features of the data are shown in Tab 2.

Tab 2 Description and analysis of physical activity

2.3.1.2 Measure method of independent variable

This study will study the diversity of activity in three aspects: the diversity of PA type, the diversity of PA time, and the diversity of participants’ class. The measuring method is shown in Tab 3.

Tab 3 measure method of slow-physical activity diversity

We overlap the PA route data with the land use data in geospatial. Since the starting point or end point of each activity route can give information on residential area, the socioeconomic status of the participants can be obtained according to the housing price of the residential area（Li & Zhang, 2008）.

The diversity of the participants’ class is according to the housing price of the activity start and end point. The information about housing price is deduced from the housing price-income ratio. The ratio of housing price to income is the ratio of the selling price of commercial housing to the annual disposable income of urban households. The formula is:

HPR=my/n

where

According to the existing literature and actual research, this paper divides the middle-income class and the high-income class according to the ownership of a 80 m² house, divides the middle-low-income class and the low-income class according to the ownership of a 60 m² house and calculates the affordable housing prices of different income classes in Shenzhen.

2.3.2 Dependent variable

The built environment indicators are divided into two categories: general index and particular index. The selection of general indicators is mainly based on the 5D model of environmental impact on resident non-motor vehicle travel. Particular indicators are useful environmental indicators selected for three different independent variables by analyzing the diversity characteristics of continuous PA. The indicators and measurement methods are shown in Error: Reference source not found.

2.3.3 Control variables

From the previous literature review and description of the PA in Shenzhen, it can be found that time factors such as if it is a working day have a particular impact on the distribution of diversity of moderate PA, so this paper takes it as a control variable. Whether it is a holiday or not (1:0) and whether it is spring or summer (1:0) is taken as control variables and assigned to this dummy variable.

2.4 Data merging and processing

The geocoded PA data, the built environment data, and housing price data were overlaid in ArcGIS 10 with the Shenzhen local coordinate system to form the database for this study.

2.5 Statistical analysis

In this paper, Statistical Product and Service Solutions (SPSS) were used for statistical and Regression analysis to establish regression models on the diversity of PA type, the diversity of PA time, and the diversity of participants’ class, and to explore the association between the diversity of moderate PA and the influencing factors. The collinearity diagnostics among the independent and control variables were done before modeling, and the result showed there

was no high association among the variables (VIF<10).

Tab 4 selection and measurement of environmental indicators

3. Results

3.1 Descriptive characteristics of the physical activity

3.1.1 The diversity of activity type

The amount of the three types of activities shows a spatial aggregation in the southeast of Shenzhen. People are doing walking activities mainly around their residential blocks and communities; there is no activity beyond its region. Jogging route is similar to the route of walking but the former is longer than the latter, and the jogging routes have a higher repetition frequency than walking. Some of the activity occurs in multiple regions. Compared with the other two activities, the spatial distribution of cycling activities is more balance (Fig. 2).

Fig. 2 Distribution of three types of activities

In this paper, the the diversity of PA type refers to the mixture level of walking, jogging and cycling in the spatial grid of Shenzhen. The PA type in Nanshan, Futian, and Luohu regions are more diverse than other regions (Fig. 3).

Fig. 3 Distribution of the diversity of PA type

3.1.2 The diversity of activity time

The diversity of PA time refers to the time of a day when people do the physical activities. The study divides the whole day into four periods: early morning (0:00-6:00), morning (6:00-12:00), afternoon (12:00-18:00) and evening (18:00-24:00). People prefer Cycling in the evening, prefer walking or jogging in the morning and evening. About 45% of the activities took place in the evening. The time distribution is shown in Table 5.

Tab 5 Temporal distribution of moderate physical activity

Fig.4 shows the distribution of PA in different time, it can be found that the activities with long distance and duration are mostly in the morning and evening, and most of the activity trajectories are linear. Compared with other periods, the activity trajectories are similar to the trace in the morning, but the amount of the activity fluctuated in all regions.

Fig. 4Time distribution of PA

According to the spatial distribution of PA in different time and the spatial distribution of time diversity of all kinds of PA, it can be seen that the time diversity is higher in places where there are activities in the early morning. The diversity of PA time was relatively higher in places with higher overlap between morning activities and evening activities (Fig 5).

Fig. 5 Distribution of the diversity of PA time

3.1.3 The diversity of PA participants’ class

According to the PA data, the middle-income class takes the most percentage of all the activities, accounting for 55% of the total. The proportion of low-income class and lower and middle class is relatively high in running activity (Table 6).

Tab. 6 Distribution of PA among different social classes

It can be seen that the social class of participant in Nanshan, Futian, Luohu, and Longhua regions are more diverse than in other regions. The number of the middle-income class in former regions is relatively more plentiful than later regions.

Fig. 6 Distribution of PA participants from different social classes

There are relatively more low-income class in Longgang, Baoan, and Guangming regions, while there are more middle-income class in Nanshan region, Luohu and Longhua region.

Fig. 7 Distribution of PA of participant social class

3.1.4 Comprehensive diversity

The comprehensive diversity of PA in Shenzhen can be obtained by spatially sum the values of the diversity of PA type, the diversity of PA time and the diversity of PA participants’ class. type-time diversity represents sustained and abundant PA, while type-social class diversity represents equitable and abundant PA. Different types of diversity have different demands on the place. By presenting the comprehensive diversity in geographic space, we can see that the support of Shenzhen public space for various diversity is different.

The public space can be divided into different characteristic venues. Professional venues with regular service distribute more homogeneously in the whole city of Shenzhen. Multi-functional venues with full-time service are relatively more, but most of them distribute in developed areas, such as Nanshan region and Futian region. Moreover, the multi-functional sports venues which servicefull-time are mostly located along the Shenzhen Bay in the southern coast of Shenzhen.

Fig.8 Characteristics and distribution of various activity places

3.2 Model and result

The results of the regression model are shown in Tab. 7. There is a positive association between the diversity of PA type and greenway density, interest point density of scenic spots, historic sites, sports facilities, road density and bus station density. At the same time, there is a significant negative association between the diversity of PA type and residential density and land use mixture.

The density of shopping and catering interest points and the level of public safety showed a significant positive association with the diversity of PA time. Street density, bus station density and the proportion of green space were positively correlated with the diversity of PA time, while the proportion of NDVI and the commercial land was negatively correlated with the diversity of PA time.

The street density, the density bus stations density, the proportion of green space, the density of interest points of public service facilities are positively correlated with the diversity of PA participants’ class.

Tab 7 Regression models and results

4. Discussion

4.1 Accessibility is the primary condition for the three types of PAs aggregation

According to the regression model, the three categories of diversity are positively correlated with street density and bus station density. Intensive street network and completely public transport system can add the possibility to arrive the destination for the people from different social classes. Besides, a well-developed transportation system can ensure the demand for different time periods, thus attract diverse activities. Therefore, the accessibility of activity spaces can significantly support all the social class to do PA at different times of the day.

4.2 A safe and comfortable environment contributes to the diversity of PA time and the diversity of PA participants’ class

The density of green space supports the diversity of PA time and participant’s class. The reason is that most of the green space is particularly designed for multiple PAs. Moreover, most of the parks in Shenzhen are open to the public all day, so residents from different classes can do activities any time they want. However, the design of parks are mostly aimed at daily activities such as walking and running, and less attention is paid to cycling which needs professional requirements, that is why there is no association between the proportion of green space and the diversity of PA type. The level of public service and the diversity of community types stand for the cohesion of the city. In open space with well-equipped and diverse community types, no matter what social class can enjoy a comfortable and good outdoor environment, and the diversity of residential types can also promote the harmony of the outdoor open spaces and the community.

Also, the diversity of PA time is closely related to the outdoor activities at night, and night-time activities have a higher demand for safety. Food and beverage shopping places often attract people to gather at night and do the activities around them because of their long open-time and lighting characteristics. On the contrary, the places with high green coverage have low visibility at night, which may give people a sense of insecurity and is not conducive to supporting night activities.

4.3 Open and well-equipped facilities could support diverse types of PA

The purpose of constructing greenway network and sports facilities are to promoting outdoor activities. Because of the planning and design, the demand for outdoor activities is considered carefully so that it can support a variety of outdoor activities. Similarly, Shenzhen's scenic spots are mostly artifactual landscapes. There is a uniform standard for the establishment of service facilities and infrastructure, which can support a variety of activities. However, the high degree of land use mixture indicates that the number of potential destinations increases, which will also cause crowds to gather. In this case, limited space will limit the diversity of PA type. Similarly, high population density will result in the reduction of usable public space, which will make the activity space narrow and the movable path complex, and limit the jogging and cycling activities which require high fluency. Therefore, in order to promote the various types of activities, the open space should firstly ensure that it is open and unobstructed, and ensure the activities have enough space and do not conflict with each other; secondly, the open space should provide professional services for different activities, such as pedestrian tracks, plastic runways, and bicycle parking areas.

5. Conclusion

In recent years, great progress has been made in the study of human settlements and public health. More and more scholars have begun to explore the value of built environment in improving public health. The built environment is an important carrier of residents' behavior and urban planning, and the level of PA is one of the important indicators to mediate built environment and public health. Diversified PAs can not only promote residents' health, but also promote residents' communication, improve social harmony and cohesion. To constructing built environment that can support diverse activities, the accessibility of public space, size of parks, facility distribution and the inclusiveness of public space are particularly important. This article aimed at providing reference for future urban planning and construction by constructing the diverse diversity of activity, at the same time to provide guidance for build a healthy city.

Reference

Boer R., Zheng Y., Overton A., et al. (2007). Neighborhood design and walking trips in ten U.S. metropolitan areas. American Journal of Preventive Medicine. 32（4）:298-304.

Cervero R, Kockelman K. (1997). Travel demand and the 3Ds: density, diversity, and design. Transportation Research Part D Transport & Environment, 2(3): 199-219.

Chen P.J., Weng X.Q., Lin W.T. (2014) Studies on built environment of physical activity promotion: a co-action between multidis-ciplinary and cross –sections. Sports & Science (1) :22-29

Feng J.X., Yang Z.S., (2015). Factors influencing travel behavior of urban elderly people in Nanjing. Progress in Geography, 34(12):1598-1608.

Forsyth A., Oakes J.M., Schmitz K.H., et al. (2007). Does residential density increase walking and other physical activity. Urban Studies. 44: 679-697.

Frank L.D. (2000). Land use and transportation interaction. Journal of Planning Education & Research, 20(1): 6-22.

Heinrich K.M., Lee R.E., Suminski R.R., et al. (2007). Associations Between the Built Environment and Physical Activity in Public Housing Residents. International Journal of Behavioral Nutrition and Physical Activity, 4（1）: 1-9.

Hou N., Popkin B.M., Jacobs D., et al. (2010). Longitudinal Associations Between Neighborhood-level Street Network with Walking, Bicycling and Jogging: The CARDIA Study. Health & Place, 16(6): 1206-1215.

Inoue S., Murase N., Shimomitsu T., et al. (2009). Association of physical activity and neighborhood environment among japanese adults. Preventive Medicine, 48（4）: 321-325.

Inoue S., Ohya Y., Odagiri Y., et al. (2010). Association between perceived neighborhood environment and walking among adults in 4 cities in japan. Epidemiol, 20(4): 277-286．

Jane Jacobs. The Death and Life of Great American Cities[M]. New York: Random House, 1961.

Jiang H.Y., Zhou C.S., Xiao R.B. (2010). Spatial differentiation and social equity of public parks in Guangzhou. City Planning Review. 34(4):43-48.

Jiang L., Yang D.F. (2012) A preliminary study on the quantitative evaluation method of urban street vitality .China urban planning annual conference proceedings.pp:21-24.

Karusisi N., Bean K., Oppert J.M., et al.(2012). Multiple dimensions of residential environments, neighborhood experiences, and jogging behavior in the record study. Preventive Medicine. 55（1）: 50-55.

Kelly C.M., Schootman M., Baker E.A., et al. (2007). The Association of Sidewalk Walkability and Physical Disorder with Area-level Race and Poverty. Journal of Epidemiology & Community Health, 61（11）: 978-983.

Lee R.E., Cubbin C., Winkleby M. (2007) Contribution of neighborhood socioeconomic status and physical activity resources to physical activity among women. Journal of Epidemiology & Community Health, 61(10):882-890.

Li L.Y., Du Q.Y., Weng M., et al. (2008) Spatial grid division of city based on domain model. National conference on cartography and GIS. pp: 96-99.

Li P.L., Zhang Y. (2008) The size, identity and social attitudes of China's middle class. Society. 28 (2):1-19.

Lin X.B., Yang J.W. (2015). Built environment and public health review and planning in north American metropolitan areas, Planners, (6):12-19.

Liu K., Siu K. W. M., Gong Y.X., et al. (2016). Where Do Networks Really Work? The Effects of the Shenzhen Greenway Network on Supporting Physical Activities. Landscape & Urban Planning, 152: 49-58.

Liu Y.Y., Tan S.H., Song S.S., et al. (2016). Research on the residential green environment enhancing the health of population. Architecture & Culture, (8):188-190.

Long Y., Zhou G. (2016). Quantitative evaluation on street vibrancy and its impact factors: a case study of Chengdu. New Architecture. (1): 52-57.

Lu F.D., Tan S.H. (2015) Built environment's influence on physical activity: review and thought. Urban Planning International. 30 (2):62-70.

Orleans C.T., Kraft M.K., Marx J.F., et al. (2003). Why are some neighborhoods active and others not? Charting a new course for research on the policy and environmental determinants of physical activity. Annals of Behavioral Medicine, 25(2): 77-9.

Pan C.Y., Jin J.B., Liu Y., et al. (2007). The Nature of Urban Diversity and Its Application in Urban Planning. HENAN SCIENCES. 25 (5) :858-861.

Rundle A.G., Sheehan D.M., Quinn J.W., et al. (2015) Using GPS data to study neighborhood walkability and physical activity. American Journal of Preventive Medicine, 50(3): 65-72.

Wang D., Han X. (2012).Children’s outdoor physical activities in Beijing “Urban-Village” neighborhood subjected environment factors: a case study on Da youzhuang and Sao ziying neighborhood. Acta Scientiarum Naturalium Universitatis Pekinensis. 48（5）：841-847.

Wang Y., Chau C.K., Ng W.Y., et al.(2016). A review on the effects of physical built environment attributes on enhancing walking and cycling activity levels within residential neighborhoods. Cities, 50: 1-15.

Wang Z, Lee C. (2010). Site and neighborhood environments for walking among older adults. Health & Place, 16（6）: 1268-1279.

Wang Z., Lee C.(2010). Site and Neighborhood Environments For Walking Among Older Adults. Health & Place, 16(6):1268-1279.

Weng X.Q., He X.L., Wang X.S., Lin W.T., Li D.H. (2010). Influence of urban architectural environment on resident physical activity and health: a new field of sports and health promotion research, China Sport Science, 30 (9) :3-11

Xu F., Li J., Liang Y., et al. (2010). Associations of residential density with adolescents physical activity in a rapidly urbanizing area of mainland China. Journal of Urban Health, 87(1):44-53.

Xu L.Q., Liu N., Lu J.W. (2015) The impact of density, ratio and microcosmic quality of public space on urban vitality: microscopic observation of shanghai rail transit station areas. New Architecture. (4): 21-26.

Zapata Diomedi B, Veerman J L. (2016). The association between built environment features and physical activity in the australian context: a synthesis of the literature, BMC Public Health, 16(1): 1-10.

Zhang C.Y., Zhang G., Zhou H.Y.(2017). The analysis and influence mechanism research of urban vigorous space based on multiple big data: a case study on the partial area of central Hangzhou. Architecture & Culture, (9):183-187.

Zhang Y., Chen L., Liu H. (2010). Impact of physical activity-related environment on healthJournal of Environment and Health. 27 (2):165-168.

Zhang Z. (2008) Research on urban mixed-income housing. Doctoral dissertation. Harbin Institute of Technology.

Zhou L., Wang T., Ma N. (2008). Urban public green space accessibility research based on GIS——take Hankou of Wuhan as an example. Yunnan Geographic Environment Research.20 (4) :11-15

Zhou R.N., Li Y., Fu H. (2012). Review of residential neighborhood environment influence on residents physical activity,Chinese Journal of Health Education, 28 (9) :769-771

Zou Q.Y. (2015). Research on the different social strata leisure sports in Henan province. Doctoral dissertation. Henan Normal University in Partial Fulfillment of the Requirements.