Skip to content

Advertisement

International Journal of Behavioral Nutrition and Physical Activity

What do you think about BMC? Take part in

Open Access

Designing healthy communities: creating evidence on metrics for built environment features associated with walkable neighbourhood activity centres

  • Lucy Dubrelle Gunn1Email author,
  • Suzanne Mavoa2,
  • Claire Boulangé1,
  • Paula Hooper3,
  • Anne Kavanagh4 and
  • Billie Giles-Corti1
International Journal of Behavioral Nutrition and Physical Activity201714:164

https://doi.org/10.1186/s12966-017-0621-9

Received: 23 February 2017

Accepted: 22 November 2017

Published: 4 December 2017

Abstract

Background

Evidence-based metrics are needed to inform urban policy to create healthy walkable communities. Most active living research has developed metrics of the environment around residential addresses, ignoring other important walking locations. Therefore, this study examined: metrics for built environment features surrounding local shopping centres, (known in Melbourne, Australia as neighbourhood activity centres (NACs) which are typically anchored by a supermarket); the association between NACs and transport walking; and, policy compliance for supermarket provision.

Methods

In this observational study, cluster analysis was used to categorize 534 NACs in Melbourne, Australia by their built environment features. The NACS were linked to eligible Victorian Integrated Survey of Travel Activity 2009-2010 (VISTA) survey participants (n=19,984). Adjusted multilevel logistic regressions estimated associations between each cluster typology and two outcomes of daily walking: any transport walking; and, any ‘neighbourhood’ transport walking. Distance between residential dwellings and closest NAC was assessed to evaluate compliance with local planning policy on supermarket locations.

Results

Metrics for 19 built environment features were estimated and three NAC clusters associated with walkability were identified. NACs with significantly higher street connectivity (mean:161, SD:20), destination diversity (mean:16, SD:0.4); and net residential density (mean:77, SD:65) were interpreted as being ‘highly walkable’ when compared with ‘low walkable’ NACs, which had lower street connectivity (mean:57, SD:15); destination diversity (mean:11, SD:3); and net residential density (mean:10, SD:3). The odds of any daily transport walking was 5.85 times higher (95% CI: 4.22, 8.11), and for any ‘neighborhood’ transport walking 8.66 (95% CI: 5.89, 12.72) times higher, for residents whose closest NAC was highly walkable compared with those living near low walkable NACs. Only highly walkable NACs met the policy requirement that residents live within 1km of a local supermarket.

Conclusions

Built environment features surrounding NACs must reach certain levels to encourage walking and deliver walkable communities. Research and metrics about the type and quantity of built environment features around both walking trip origins and destinations is needed to inform urban planning policies and urban design guidelines.

Keywords

Transport walkingPlanning policyBuilt environmentUrban designNeighbourhood activity/town centreCluster analysisLand use mixGeographic information systems

Background

Active transport – including walking for transport– [14] is both health-promoting and supports sustainable living [5, 6]. A growing body of international evidence [710], shows the built environment plays an important role in creating pedestrian-friendly neighbourhoods that promote walking and reduce chronic disease risk factors.

A number of built environment features are consistently shown to facilitate transport walking around residential homes, which are the origins of many walking trips. These include: highly connected streets, high population density, mixed land use and good access to destinations and transit, and sidewalk provision [1115]. However, to inform planning policy and urban design guidelines, policy-makers and urban designers require specific information on the types, quantities and mix of built environment features that influence walking [3, 1621]. Moreover, most of the research to date has focused on the home environment [22], with less research focused on the end points of those trips.

Of all the built environment features that promote walking, it is destinations that are most important because they provide a ‘reason to walk’ and an end-point for walking trips [23]. Destinations and destination diversity are repeatedly shown in active living research studies to be positively associated with transport walking [10, 2429]. Yet, unlike transport academics [30], active living researchers rarely study the built environment features surrounding destinations when undertaking travel mode studies.

In local communities, destinations are often co-located and concentrated at local shopping centres, known in Melbourne, Australia as ‘neighbourhood activity centres’ (NACs) [25, 31, 32]. NACs function as a community focal point and enable local living by providing a variety of retail shops, services and destinations. Their distribution across cities are important because they facilitate the creation of the ‘20- (or 30-minute) city,’ a concept that is drawing growing global interest [3234]. However, to attract pedestrians the built environment surrounding NACs must support walking [33, 35, 36].

Understanding the type, quantities and mix of built environment features and their influence on walking trips is therefore crucial for designing and retrofitting NACs to be pedestrian-friendly. One method to evaluate this is cluster analysis, which groups built environment features to help identify walkable neighbourhood environments [3, 4, 25, 3641]. Several authors have modeled associations between cluster type (treated as an exposure variable) and walking, and have found that more walkable neighbourhoods (as defined by cluster analysis) are typically associated with increased odds of walking [3, 4, 25]. Most research using cluster analysis has focused on home neighbourhoods [4, 25, 37, 40], with few studies, if any, providing metrics on the type, quantities and mix of built environment features that encourage walking or consider destinations, such as NACs, that could inform relevant urban design guidelines [36].

In Melbourne, Australia, precinct structure plans typically allow for town centres or NACs to be anchored by supermarkets. Furthermore, Melbourne’s planning policy stipulates that “80-90% of households should be within 1km of a town centre of sufficient size to allow for provision of a supermarket” [42]. However, unless policy implementation is evaluated and measured on the ground [43] it remains unclear whether this policy requirement is being met. Indeed, few studies measure urban design policy implementation in Australia [3, 44]. Nevertheless, this type of evidence is useful to assess whether NACs have the potential to facilitate transport walking [45]; and, could help achieve the desired 20-minute city in Melbourne.

The aims of this paper were to: (1) investigate the type, quantity and mix of a broad range of built environment features to create a walkability typology of NACs across Melbourne, Australia; (2) examine whether built environment features surrounding NACs are associated with local residents’ transport walking; (3) assess the spatial distribution of NACs in relation to walkability; and (4) assess compliance with local planning policy about the location of supermarkets in relation to residential dwellings.

Methods

Individual-level data

Individual level data from the Victorian Integrated Survey of Travel Activity (VISTA) 2009-2010 were used [46]. VISTA is a repeat cross-sectional survey of one-day travel behavior administered by the Victorian Department of Economic Development, Jobs, Transport, and Resources.

Data for urban areas of metropolitan Melbourne were extracted for this study for 21,664 adults aged 18 years and over. VISTA follows a multi-stage stratified sampling process. First, a stratified random sample was taken across eight regions in Melbourne, and within each region several Census Collector Districts (CCD) were selected followed by sampling individual households within CCDs. CCDs cover an average of 225 dwellings. Due to low response rates in some areas, additional households were selected to meet response rate requirements. The survey asks questions on one day of travel behavior for each individual within each household. Further information on the VISTA survey can be found in online Additional file 1: Appendix 1 VISTA Sample.

Activity centre data

The NAC dataset was defined as all commercially zoned property boundaries within an 800m street network buffer of selected supermarkets. An 800m buffer distance was used here, since several researchers have found associations between built environment correlates and physical activity measures at this distance [4749]. In particular, Manaugh et al (2011) found associations with Walkscore and a Walkability Index and home-based shopping trips at 800m [49] and Gunn et al (2017) found increased odds of transport walking for those located within 800m of local food outlets including supermarkets, cafés/takeaway stores, and small food stores, which are typically co-located within NACs [48].

Property data were sourced from VicMap Property [50]. Street centrelines were sourced from VicMap Transport [51] and freeways and freeway on-ramps were removed to create a walkable road dataset from which the 800m street network buffers around NACs were created.

Geocoded supermarket data were obtained from a commercial provider [52] and spatially mapped using Arc GIS 10.2. This data set was validated against the state’s strategic planning framework which identifies major NACs across Melbourne (i.e., Plan Melbourne [33]) through which a number of additional supermarkets were identified and manually geocoded. Data for commercial zones were obtained from VicMap Planning [50]. This was overlaid onto the supermarket layer, to ensure that identified supermarkets were associated with commercial zones with other retail stores. Manual checking of the NAC layer was conducted using Google Maps. Supermarkets not placed within a commercial zone were removed from the dataset. Since many NACs contain more than one supermarket, duplicate supermarket addresses and duplicate NACs were removed from the dataset, however information on how many different supermarkets existed in each NAC was retained as a supermarket diversity variable.

Dependent variables

Two dependent variables were developed using VISTA daily transport walking trip data. First, due to the non-normal distribution of VISTA walking data, any transport walking was defined as all daily transport walking trips dichotomized into ‘one or more walking trips’ and ‘no walking trips’[44, 53]. Second, to capture local transport walking, any neighbourhood transport walking was defined as daily walking trips beginning and ending within a 1.6km street network buffer of a participants’ home. This data was dichotomized into ‘one or more walking trips within the neighbourhood’ and ‘no walking trips in the neighbourhood’ (see Fig. 1).
Fig. 1

Dependent variable definitions for transport walking trips. a Any transport walking. b Any neighbourhood transport walking

Independent variables

Built environment

Following the work of Hooper et al (2014), metrics for 19 built environment features were defined for three domains: community design, movement network, and lot layout [43].

Community design was measured using the presence and quantity of a variety of destinations with accessibility to these destinations measured using the number of transport stops and pedshed ratios (i.e.,the ratio of area within 800m street network buffer to the area within 800m Euclidean buffer). Based on planning guidelines, Hooper et al (2014) also define the movement network made up of design features such as street connectivity, block perimeters and the number of cul de sacs [54] which determines the directness of pedestrian routes and the proximity of local destinations. The lot layout domain includes housing diversity and net residential density [43], which determines the number of residents living in close proximity to NACs and hence, supports the viability of retail shops and transit services. Built environment features were calculated within an 800m street network buffer of the NAC using GIS methods and ArcGIS 10.2 software (see Table 1).
Table 1

Built environment variables calculated within 800m of a supermarket

Variables and definitions

Community Design

 Pedshed: ratio of area within 800m street network buffer to the area within 800m Euclidean buffer

 Number of community resources: post offices, community centres, child care centres, libraries

 Number of small food stores: butcher, green grocers, convenience stores

 Number of other retail stores: banks, pharmacy, petrol station, newsagent

 Number of supermarkets: includes major and minor supermarkets

 Supermarket diversity: number of different major supermarkets (0-4)

 Number of transport stops: buses, trams, train

 Transport diversity: number of different types of transport (0-5)

 Destination diversity: number of different individual destination types (0-16)

Movement Network

 Street connectivity: number of ≥3 way intersections

 Cul de sacs: number of cul de sacs

 Cul de sac segments ≤120m long: number of cul de sac segments ≤120m long

 Connected node ratio: number of ≥3 way intersections ÷ all intersections including cul de sacs

 Disconnected node ratio: number of cul de sacs ÷ all intersections including cul de sacs

 Mean block perimeter (m)

 Walkable block ratio: number of blocks ≤620m perimeter ÷ total number of blocks

 Traffic exposure ratio: length of low traffic roads ÷ length of low and high traffic roads

Lot Layout

 Housing diversity: number of different housing types (0-8)

 Net residential density: number of commercial dwellings + number of residential dwellings ÷ commercial and residential area

Confounders

A number of confounders likely to affect transport walking trips undertaken by participants were identified from the literature including: age, sex, studying (yes, no), occupation (unemployed, casual work, part-time, full-time), household type (couple no children, couple with children, other household structure including single parent, sole person, and other household structure undefined), income (<$650, $650-$1099, $1200-$1949, $1950-$2499, $2500), own a motor vehicle (do not own a motor vehicle, own a motor vehicle), and area level disadvantage derived from the Australian Bureau of Statistics Socio-Economic Index for Areas, Index of Relative Social Disadvantage (SEIFA IRSD) based on deciles (high: 1-4, medium: 5-7, low: 8-10) [44].

Statistical analysis

Cluster analysis of neighbourhood activity centres

Cluster analysis creates homogenous clusters of built environment variables whilst ensuring that each cluster is distinct from the others [55]. It was used to identify NAC typologies based on the built environment features listed in Table 1.

Specifically, a two-stage cluster analysis was undertaken. K-means clustering was used since it performs better in the presence of outliers and irrelevant clustering variables [55], however the number of clusters must be specified a priori. To help establish this, in the first stage, hierarchical clustering and the Calinski-Harabasz pseudo F statistic were estimated. The number of clusters was identified from large changes in the error coefficients between stages in the agglomeration schedule of the hierarchical cluster analysis, and from identifying the largest Calinski-Harabasz pseudo F statistic for each cluster solution.

The final K-means cluster solution was chosen based on interpretability of the average values of the built environment features. This was undertaken to identify cluster types relating to walkability. The final NAC cluster solution was mapped to examine the spatial location of the NAC cluster typologies across metropolitan Melbourne.

ANOVA and Tukey post-hoc tests were undertaken on the final K-means cluster solution to establish whether there were significant differences for each built environment feature by cluster type. Results are presented as a cross tabulation of the built environment features by cluster type. The tests and the cross tabulation cluster solution indicate which built environment features are different between NAC cluster types and by ‘how much’. This information is useful for developing and monitoring evidence-based policy and design standards for new urban areas and NACs, and for retrofitting NAC areas to increase walkability.

All clustering was undertaken on standardized built environment features to remove the effect of differing scales.

Measuring associations with transport walking trips and NAC cluster types

Several participants had no measure of area level disadvantage (n=155) and a further 7% of the sample were not associated with a NAC due to living on the extreme periphery of the city. Participants with these observations were removed from the sample with all remaining analyses conducted on complete cases within urban areas of Greater Melbourne on a sample of 19,984 observations. Cross tabulations between participant’s individual level variables and the NAC cluster types were undertaken.

Multilevel logistic regression models were estimated to measure associations between the NAC cluster type included as dummy variables with the LW NACs as the reference, and any transport and any neighbourhood transport walking trips respectively. Models were adjusted for confounders and allowed for clustering of individuals within households within the geography of statistical area level 2 (SA2). CCDs are clustered within SA2 areas. SA2 areas typically cover an average population of 10,000 people and allow sufficient coverage of clustered households and their 1.6km street network buffers. Odds ratios and 95% confidence intervals are reported. All analyses were undertaken using SPSS and Stata IC 13.0.

Measuring policy implementation of NAC access

A descriptive analysis was undertaken to understand NAC access of VISTA survey participants. The closest NAC was identified for each participant’s residential location using the street network. The distance to the closest NAC and the percentage of participants who lived within various threshold distances were calculated. The first distance threshold of 1km was based on current policy guidelines in Victoria that stipulate that “80-90% of households should be within 1km of a town centre of sufficient size to allow for provision of a supermarket” [42]. The second distance threshold was based on recent empirical findings suggesting that locating local food outlets at distances of up to 800m may influence transport walking trips [48]. The third threshold distance of 1.6km was based on a commonly used distance for determining walkable neighbourhoods which is said to be associated with achieving a brisk 15 minute walk [4, 56].

Results

Cluster analysis

Results from the hierarchical cluster analysis confirmed that three clusters were appropriate for the NAC data. Table 2 and Fig. 2 present summary statistics and a map for the three cluster solution respectively.
Table 2

Descriptive Statistics for the three cluster solution for NACs

Cluster:

Overall: n=534

High (HW) : n=48

Moderate (MW): n=267

Low (LW): n=219

Built Environment Variables

Mean

SDa

Mean

SD

Mean

SD

Mean

SD

Community Design

 Pedshed

0.53

0.10

0.63 **

0.04

0.57 **

0.07

0.46**

0.09

 Community resources

11.03

7.75

27.19 **

6.28

12.42 **

6.06

5.79 **

2.47

 Small food stores

16.54

18.63

62.63 **

22.32

17.55 **

10.44

5.20 **

3.19

 Other retail

21.37

32.33

101.77 **

60.36

18.97 **

9.23

6.67 **

4.35

 Supermarkets

3.64

3.53

12.08 **

5.53

3.55 **

1.79

1.89 **

1.04

 Supermarket diversity

1.17

0.47

1.02

0.14

1.11

0.38

1.26 **

0.59

 Transport stops

70.22

48.92

183.52 **

61.00

76.51 **

26.18

37.71 **

16.59

 Transport diversity

2.13

1.04

3.71 **

1.05

2.44**

0.84

1.40 **

0.56

 Destination diversity

12.80

2.92

15.85 **

0.36

14.14 **

1.63

10.49 **

2.82

Movement Network

 Street connectivity

73.74

33.94

160.77 **

20.20

71.66 **

21.59

57.21 **

14.54

 Cul de sacs

78.43

60.02

217.38 **

75.74

51.30 **

25.66

81.06 **

40.09

 Cul de sac segments≤120m long

64.53

59.15

210.19 **

74.75

40.70 **

23.70

61.66 **

35.91

 Connected node ratio

0.19

0.08

0.21 **

0.05

0.13 **

0.04

0.27 **

0.06

 Disconnected node ratio

1.05

0.52

1.34

0.39

0.71 **

0.27

1.39

0.52

 Mean block perimeter

1428.44

1023.18

691.20

218.72

941.74

381.67

2183.40 **

1178.63

 Walkable block ratio

0.52

0.20

0.78 **

0.05

0.60 **

0.14

0.37 **

0.15

 Traffic exposure ratio

0.81

0.08

0.73 **

0.07

0.81

0.07

0.83

0.08

Lot Layout

 Housing diversity

6.55

1.55

6.48

1.66

6.54

1.55

6.57

1.52

 Net residential density

20.96

28.55

76.99**

65.31

19.82 **

14.38

10.07 **

2.97

** Significantly different between remaining cluster types based on Tukey post-hoc tests with p<0.01

SD Standard deviation

Fig. 2

Map of metropolitan Melbourne with neighbourhood activity centres (NACs) displayed by walkability

The first cluster identified had built environment features that were more pedestrian-friendly or ‘highly walkable’ (HW) in terms of community design (i.e., higher pedshed ratios, higher mean values for all destinations, and higher transport and destination diversity scores); the movement network (i.e., higher street connectivity, higher walkable block ratios) and lot layout (i.e., higher net residential densities). Despite being HW, these NACs had higher mean numbers of cul de sacs, and cul de sac segments≤120m long and almost no supermarket diversity.

The second cluster characterized ‘moderately walkable’ NACs (MW) with moderate values for most community design, movement network and lot layout domains. The MW NACs had the lowest values for cul de sacs and cul de sac segments≤120m long, connected and disconnected node ratios. However, destination diversity near MW NACs remained high and was similar to the HW NACs; and, they also had similar mean values to the HW NACs for the mean block perimeter. Traffic exposure ratios in MW NACs were similar to the final cluster.

The final cluster was characterized as ‘low walkable’ (LW) because NACs in this cluster had the lowest mean values for most of the three aforementioned domains. Table 2 shows that the low walkability of these NACs was also reflected in some of the remaining movement network variables (i.e., high cul de sacs, cul de sac segments ≤120m long, large mean block perimeter, and high traffic exposure ratio). Conversely, the LW NACs had the highest mean values for connected node ratios, but a similar mean value to HW NACs for disconnected node ratios. Housing diversity was similar for all three clusters.

Three case studies were chosen to illustrate differences in the built environment features of each NAC type. Aerial satellite imagery, street frontage views and summary statistics for each case study can be found in the online Additional file 2: Appendix 2 Case Studies.

Location of NACs

Figure 2 shows the spatial distribution of NACs by cluster type. HW NACs are generally located close to the Central Business District (CBD) with the remaining cluster types distributed according to distance from the CBD. A small number of MW NACs were located in outer urban areas, and a small number of LW NACs were located in middle suburbs.

Socio-demographic profiles and walking behavior

Table 3 presents the socio-demographic profile of VISTA participants. Those living near HW NACs, were more likely to be studying and living in other household structures. They were also more likely to be working full-time and have incomes exceeding $2500 per week, and are more likely to live in areas with low area-level disadvantage. While there were high proportions of motor vehicle owners in all three clusters, this was lower (81%) for those living near HW NACs.
Table 3

Socio-demographic profiles of survey participants across neighbourhood activity centre types

Covariate

  

Walkability

  

Overall

(n=19,984)

N (%)

High (HW)

(n=757)

N (%)

Moderate (MW)

(n=9,259)

N (%)

Low (LW)

(n=9,968)

N (%)

Sex

Male

9,516 (47.6)

361 (47.7)

4,344 (46.9)

4,811 (48.3)

Female

10,468 (52.4)

396 (52.3)

4,915 (53.1)

5.157 (51.7)

Agea

 

46.34 (17.1)

43.13 (15.8)

46.85 (17.6)

46.10 (16.6)

Studying

Yes

2,179 (10.9)

128 (16.9)

1,097 (11.9)

954 (9.6) *

No

17,805 (89.1)

629 (83.1)

8,162 (88.2)

9.014 (90.4)

Occupation

Unemployed

6,225 (31.2)

194 (25.6)

2,932 (31.7)

3,099 (31.1)*

Casual work

1,522 ( 7.6)

54 ( 7.1)

751 (8.1)

717 (7.2)

Part-time

3,053 (15.3)

115 (15.2)

1,405 (15.2)

1,533 (15.1)

Full-time

9,184 (46.0)

394 (52.1)

4,171 (45.1)

4,619 (46.3)

Household Structure

Couple with children

7,755 (38.8)

437 (57.7)

3,867 (41.8)

3,451 (34.6) *

Couple without children

10,070 (50.1)

195 (25.8)

4,290 (46.3)

5,585 (56.0)

Other household structure

2,159 (10.8)

125 (16.5)

1,102 (11.9)

932 (9.4)

Income

<$650

2,387 (11.9)

102 (13.5)

1,194 (12.9)

1,091 (11.0) *

$650-$1099

3,227 (16.2)

105 (13.9)

1,443 (15.6)

1,679 (16.8)

$1200-$1949

3,724 (18.6)

125 (16.5)

1,584 (17.1)

2,015 (20.2)

$1950-$2499

4,357 (21.8)

132 (17.4)

1,890 (20.4)

2,335 (23.4)

$2500+

6,289 (31.5)

293 (38.7)

3,148 (34.0)

2,848 (28.6)

Own a motor vehicle

Do not own a motor

946 ( 4.7)

139 (18.4)

609 ( 6.6)

198 ( 2.00) *

Own a motor vehicle

19,038 (95.3)

618 (81.6)

8,650 (93.4)

9,770 (98.0)

Area level disadvantage b

High

5,545 (27.8)

128 (16.9)

2,535 (27.4)

2,882 (28.9) *

Medium

5,798 (29.0)

124 (16.4)

2,442 (26.4)

3,232 (32.4)

Low

8,641 (43.2)

505 (66.7)

4,282 (46.3)

3,854 (38.7)

Statistical significance between clusters assessed using chi-squared tests for categorical variables and ANOVA for continuous variables: p<0.05

Means and standard deviations presented

Based on SEIFA IRSD deciles where 1-4 indicate higher levels of area based disadvantage, 5-7 is medium and 8-10 is low

There were many similarities in the socio-demographic profile of people living near MW and LW NACs. Participants living near these two areas were similar in terms of occupation and area level disadvantage, with almost one third unemployed, and around two thirds living in areas with high or medium disadvantage. Notably, participants living near LW NACs were less likely to be studying and more likely to be a couple with children compared with participants living near MW NACs. Almost 100% of participants living near LW NACs owned a motor vehicle (98%).

A descriptive analysis of walking behavior showed that 22.1% or 4,412 participants undertook any transport walking and 16.7% or 3,339 participants undertook any neighbourhood transport walking.

Association between NAC type and transport walking trips

Table 4 presents descriptive statistics for road network distances to NACs. The mean distance for people living near HW NACs was 598.8m (SD: 335.4m). The descriptive statistics indicate that overall 76% of survey participants had their closest NAC within 1.6km; 44% were within 1km; and, 29% were within 800m. For people living near HW NACs, 88% lived within 1km, which meets Victoria’s policy requirement that “80-90% of households should be within 1km of an activity centre of sufficient size to allow for provision of a supermarket” [42].
Table 4

Descriptive statistics on the distance to the closest neighbourhood activity centre by type

Neighbourhood

Activity Centre

Obsa

Mean

SDa

Min

Max

Obs

≤1.6km

%

≤1.6km

Obs

≤ 1km

%

≤ 1km

Obs

≤ 800m

%

≤ 800m

High (HW)

757

598.8

335.4

27.0

1,826.6

754

99.6

665

87.9

548

72.4

Moderate (MW)

9,259

1,002.6

589.3

2.8

5,162.3

8001

86.4

5,368

58.0

4005

43.3

Low (LW)

9,968

1,459.6

738.9

75.1

5,042.1

6383

64.0

2,752

27.6

1736

17.4

Overall

19,984

1,215.3

708.9

2.8

5,162.3

15,138

75.8

8,785

44.0

5741

28.7

Obs Number of observations, SD Standard deviation

Table 5

Logistic regression results of any transport walking and any neighbourhood transport walking by neighbourhood activity centre type (n=19,984) a

Neighbourhood

Activity Centre

Any transport walking

OR (95% CI)b

Any neighbourhood transport walking

OR (95% CI)

Low (LW)

ref

ref

Moderate (MW)

2.28 (1.96, 2.66)***

2.75 (2.26, 3.33)***

High (HW)

5.85 (4.22, 8.11)***

8.66 (5.89, 12.72)***

*** p<0.001

Models adjusted for the following confounders: sex, age, studying, household structure, income, motor vehicle ownership, area level disadvantage. Models estimated using 3 level multilevel logistic regression according to SA2, household and individual levels

OR Odds ratios, CI Confidence intervals

Multilevel logistic regression results showed that compared with those living near LW NACs, those located near HW and MW NACs had increased odds of undertaking 1 or more transport walking trips (HW, OR: 5.85, CI: 4.22, 8.11 and MW, OR: 2.28, CI: 1.96, 2.66) and 1 or more neighbourhood transport walking trips (HW, OR: 8.66, CI: 5.89, 12.72 and MW, OR: 2.75, CI: 2.26, 3.33) (Table 5).

Discussion and conclusion

Cluster findings

Clustering NACs by their built environment features produced three clusters characterized empirically by their level of walkability consistent with previous research [4, 37]. The NAC cluster typologies were generally located in inner, middle and outer areas of Melbourne, which, although not analyzed here, relates to their period of construction from the 1800s through to the 21st century. Overall, more walkable NACs were located in older neighbourhoods.

Summary and discussion of NAC built environment cluster results

HW NACs were characterized by pedestrian-friendly community design, movement network and lot layout. However, contrary to what might be expected, in Melbourne, HW NACs also had high cul de sac values - which have been shown to detract from the walkability of an area [37, 57, 58]. This appears to be because streets in Melbourne’s inner city were designed with rear laneway access for household services and sewerage collection that now provide access to subdivided blocks and small studio style apartments. Furthermore, many inner city streets have been blocked to encourage traffic flows on nearby arterial roads instead of on residential streets. Indeed, more high traffic volume roads were found in HW NACs as evidenced by the low traffic exposure ratio. HW NACs tended to be located on arterial ‘main street’ style shopping strips that house a diversity of shops, services, and high quality public transport, which together have been associated with transport walking [23]. Typically, gridded street networks sit behind these arterial streets providing easy pedestrian access, which is consistent with the notion of transit oriented development [59].

The MW NACs retained many, but not all, of these pedestrian-friendly built environment features with the lowest values for cul de sacs and cul de sac segments≤120m long, connected and disconnected node ratios. This suggests that whilst these areas are walkable and encourage transport walking, they are also conducive to driving, with the streets retaining an open grid structure unhampered by cul de sacs.

The LW NACs had the least pedestrian-friendly design and were more spatially dispersed across the outer suburbs with greater distance to, and between NACs leading to potential inequity in access to local living destinations, employment opportunities, and public transport for residents in these areas. The summary statistics for the built environment features support the low walkability of these NACs with the curvi-linear street structure a likely contributor. The LW NACs showed large tracts of land dedicated to car parking, which has been shown to be a barrier for walking and a design that privileges motor vehicle travel as a means of getting to NACs and other destinations [6062] (see online Additional file 2: Appendix 2 Case Studies).

Finally, regardless of NAC typology there appeared to be limited differences in housing diversity with approximately 6-7 different types of housing available. To some extent, similarities in housing diversity was expected, since Melbourne is a large homogenous low density sprawling city, particularly in the middle and outer suburbs where detached housing is prominent and where infill density is only now starting to occur [63]. Policy and practice could facilitate the delivery of greater housing diversity to suit different housing requirements across the life course facilitating ageing in place [64]. Increased housing diversity is also a good way to increase dwelling density, which supports walkability by making public transport, services and retail more viable. It also helps to deliver the 20 minute city[63, 65]. However, at least in Australia, current planning policy focuses mainly on increasing densities in inner and middle areas with less emphasis on outer areas. Yet low density development in outer suburban areas not only fosters motor vehicle dependencies, but reduces access to the early delivery of public transport, shops and services. Hence these areas need the most planning and policy support,[63], since many studies now show that where we live can affect our health [6668] and access to services is a social determinant of health [69]. Furthermore, the metrics for LW NACs show low transport diversity and low numbers of transport stops compounding the disadvantage experienced in these areas, since residents have little choice but to purchase at least one motor vehicle in order to access jobs and essential services [70]. This is supported by our results, showing a correlation between area level disadvantage and the presence of either MW or LW NACs suggesting that there is some socio-spatial patterning and disparities in the provision and type of NAC across middle and outer suburbs. This finding also aligns with the lack of policy implementation which we consider further below.

Are different NAC types associated with transport walking?

Multilevel logistic regression models assessing the influence of the three NAC types on transport walking showed a significant trend of increased odds of any transport walking and any neighbourhood transport walking corresponding to the more walkable NACs. Those living near HW NACs had odds 6-9 times higher for walking for transport than those living near LW NACs. Like other authors [4, 29, 71, 72], these results suggest that built environment features have considerable influence on transport walking behavior in providing both access and destinations to walk to [27, 29, 59, 73].

Planning policy implementation of supermarket provision

Few studies objectively evaluate the implementation of urban design and planning policies [43]. Hence, we assessed the current state government policy requirement that “80-90% of households should be within 1km of a town centre of sufficient size to allow for provision of a supermarket.” We found that only households located near HW NACs met the policy requirement at a threshold of 1km (88%). Indeed, only 28% of households with LW NACS met this policy requirement. This highlights discrepancies between current policy requirements and policy implementation, especially for those living in outer suburban areas where NACs are spatially dispersed, suggesting inequitable delivery of this policy requirement across the city.

To support walking and population health there needs to be a greater focus on adherence and implementation of policy requirements [43], especially since many policies are non-binding, as is the case here. Our findings provide some evidence to support governments to strengthen policy requirements for reducing the distance between residences and NACs; and to support greater efforts in evaluating policy implementation in practice. These findings provide a “call to action” to provide well designed areas that support people and communities. This is important since it is difficult to retrofit neighbourhoods to build NACs or to provide other important infrastructure such as parks and open spaces, public transport, and schools without ensuring adequate land is allocated for these purposes in the land use planning phase of a housing development.

Strengths and Limitations

The focus on objective measures of the built environment surrounding NACs is a strength of this paper. It contributes to emergent evidence of the importance of other environmental exposures beyond where we live for influencing walking behaviors [7476], which may differ for different types of participants [77, 78]. Another strength, is the presentation of metrics of the built environment surrounding NACs. These are presented for several built environment features, which could inform future urban design guidelines. These metrics could also be used to evaluate and monitor existing and proposed developments and for retrofitting LW NACs to the standards set by MW or HW NACs where appropriate. The provision of metrics for individual built environment features and the results of the cluster analysis examining several features jointly are also strengths of this paper, since it is important to understand the types, quantities and mix of built environment features [40, 79].

However, limitations of this analysis include using the road network for GIS calculations of street connectivity and cul de sacs, which may ignore other networks accessible to pedestrians and cyclists. This may under-estimate the values for these built environment variables. Similarly, other built environment features could be added to the analysis, such as the number of transport routes, length of cycleways or, the size or the period of construction of the NAC. This may alter the cluster solution. We did not find any associations with housing diversity. It is possible that our housing diversity variable was not sensitive enough to pick-up variation in housing types. Future research may seek to refine this variable by incorporating the area of lots as well as the type and size of housing.

In this study, we used current local planning documents Plan Melbourne to define and inform the identification of NACs. The method adopted may have missed some NACs. The analysis focused on the area surrounding the NAC and measured associations with transport walking trips made by survey participants’ in their home neighbourhoods measured by a 1.6km network buffer. It is possible that survey participants’ undertook transport walking trips in other areas (e.g., near work) and to NACs other than the closest one, as was assumed in this study. We used cross-sectional data and hence, do not know the temporality of associations and consequently we could not assess whether socio-spatial disparities and differences in transport walking were driven by self-selection or social inequity or a combination of both. However, no information on neighborhood preferences was collected in the VISTA 2009-2010 survey and whilst more recent VISTA data is available a temporal mismatch between the survey and GIS data negated its use. Issues, like these, relating to deprivation amplification [66] in service provision and transportation across Melbourne are complex and are left for future work. The labeling of the cluster typologies according to levels of walkability (e.g. high, moderate, low) was empirically driven and relates to the context of Melbourne, Australia. We recommend that future research place such findings within the international context aligned to emergent evidence on thresholds for built environment variables [10]. Nevertheless, the findings here show that certain levels for the types, quantities and mix of built environment features are required to support transport walking. Finally, the dependent variables were based on one survey day measuring transport walking behavior. As such, the two definitions of transport walking used here were insufficient for determining whether or not survey participants undertook adequate physical activity.

Conclusion

This study provides evidence on the type, quantity and mix of built environment features that create walkable NACs. Results show an inequitable spatial distribution across metropolitan Melbourne, with HW NACs generally located in inner areas, and LW NACs in outer areas. Furthermore, LW NACs in outer areas were less likely to meet policy guidelines on supermarket provisioning, with these areas providing fewer active transport options and resulting in lower levels of transport walking. Overall, the results suggest that both the built environment features and dwelling density surrounding NACs play a key role in influencing travel behavior and in particular transport walking.

Features of HW NACs that encourage walking include: high street connectivity (mean:161), high destination diversity (mean:16), and high residential density (mean:77). These metrics are significantly higher than LW NACs characterized by significantly lower street connectivity (mean:57), destination diversity (mean:11), and residential density (mean:10). Identifying and providing summary statistics and metrics on a broad range of built environment features associated with increased walking is requested by decision makers [19], and will support them to develop planning policy and urban design guidelines that create healthy, more equitable and sustainable communities. This moves beyond the rhetoric of the need for more walkable communities, to supplying metrics on how this can be achieved. More research of this type in different cities is warranted.

Abbreviations

CBD: 

Central Business District

GIS: 

Geographic Information Systems

HW: 

Highly Walkable

LW: 

Low Walkable

MW: 

Moderately Walkable

NACs: 

Neighbourhood Activity Centres

VISTA: 

Victorian Integrated Survey of Travel Activity

Declarations

Acknowledgements

The Victorian Department of Economic Development, Jobs, Transport, and Resources is gratefully acknowledged for administering and providing the VISTA data. The research team is grateful for the assistance of GIS specialist Rebecca Roberts who produced the spatial map of Melbourne. We are grateful for the comments provided by the anonymous reviewers.

Funding

An NHMRC Senior Principal Research Fellow Award supports BGC (#1107672); and, LDG and PH are supported by the Centre for Research Excellence in Healthy, Liveable Communities (#1061404). SM is supported by The Australian Prevention Partnership Centre (#9100001). CB is supported by The University of Melbourne International Research Scholarship (MIRS) and the North & West Metropolitan Region Victorian Department of Health, and the NHMRC CRE in Healthy Liveable Communities (#1061404).

Availability of data and materials

The VISTA dataset can be obtained under license from the Victorian Department of Economic Development, Jobs, Transport, and Resources.

Authors’ contributions

LDG, SM, CB performed the analyses and wrote the manuscript. GIS analysis was performed by CB and SM. AK, PH and BGC collaborated in the development of the analysis and revision of the manuscript. All authors read and approved the final version of the manuscript.

Ethics approval and consent to participate

The study was approved by the University of Melbourne ethics board (Application Number 1340998.4).

Consent for publication

Not applicable.

Competing interests

The authors declare they have no actual or potential competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

(1)
Healthy Liveable Cities Group, Centre for Urban Research, College for Design and Social Context, Royal Melbourne Institute of Technology
(2)
Noncommunicable Disease Unit, Centre for Health Equity, Melbourne School of Population and Global Health, The University of Melbourne
(3)
Centre for the Built Environment and Health, School of Earth and Environment and School of Sports Science, Exercise & Health, University of Western Australia
(4)
Gender and Women’s Health Unit, Centre for Health Equity, Melbourne School of Population and Global Health, The University of Melbourne

References

  1. Frank LD, Saelens BE, Powell KE, Chapman JE. Stepping towards causation: Do built environments or neighborhood and travel preferences explain physical activity, driving, and obesity? Social Science & Medicine. 2007;65:1898–914.View ArticleGoogle Scholar
  2. Sallis JF, Bowles HR, Bauman A, Ainsworth BE, Bull FC, Craig CL, Sjöström M, De Bourdeaudhuij I, Lefevre J, Matsudo V, et al. Neighborhood Environments and Physical Activity Among Adults in 11 Countries. American Journal Of Preventive Medicine. 2009;36(6):484–90.View ArticlePubMedGoogle Scholar
  3. Hooper P, Knuiman M, Bull F, Jones E, Giles-Corti B. Are we developing walkable suburbs through urban planning policy? Identifying the mix of design requirements to optimise walking outcomes from the 'Liveable Neighbourhoods' planning policy in Perth, Western Australia. International Journal of Behavioral Nutrition & Physical Activity. 2015;12(1):1–11.View ArticleGoogle Scholar
  4. McCormack G, Friedenreich C, Sandalack BA, Giles-Corti B, Doyle-Baker PK, Shiell A. The relationship between cluster-analysis derived walkability and local recreational and transportation walking among Canadian adults. Health and Place. 2012;18:1079–87.View ArticlePubMedGoogle Scholar
  5. Hosking J, Mudu P, Dora C, Adiazola C, Welle B, Herrera S: Health and Co-benefits of Climate Change Mitigation. In.; 2011.Google Scholar
  6. Watts N, Adger NA, Annolucci P, Blackstock J, Byass P, Cai W, Chaytor S, Colbourn T, Collins M, Cooper A, et al. Health and climate change: policy responses to protect public health. Lancet. 2015;386(10006):e1861–914.View ArticleGoogle Scholar
  7. Althoff T, Sosič R, Hicks JL, King AC, Delp SL, Leskovec J. Large-scale physical activity data reveal worldwide activity inequality. Nature. 2017;Google Scholar
  8. Hirsch JA, Moore KA, Clarke PJ, Rodriguez DA, Evenson KR, Brines SJ, Zagorski MA, Diez Roux AV. Changes in the Built Environment and Changes in the Amount of Walking Over Time: Longitudinal Results From the Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol. 2014;180(8):799–809.View ArticlePubMedPubMed CentralGoogle Scholar
  9. Creatore MI, Glazier RH, Moineddin R, Fazli GS, Johns A, Gozdyra P, Matheson FI, Kaufman-Shriqui V, Rosella LC, Manuel DG, et al. Association of Neighborhood Walkability With Change in Overweight, Obesity, and Diabetes. JAMA. 2016;315(20):2211–20.View ArticlePubMedGoogle Scholar
  10. Sallis JF, Cerin E, Conway TL, Adams MA, Frank LD, Pratt M, Salvo D, Schipperijn J, Smith G, Cain KL, et al. Articles: Physical activity in relation to urban environments in 14 cities worldwide: a cross-sectional study. Lancet. 2016;387:2207–17.View ArticlePubMedGoogle Scholar
  11. McCormack G, Shiell A. search of causality: a systematic review of the relationship between the built environment and physical activity among adults. International Journal of Behavioral Nutrition and Physical Activity. 2011;8:125.View ArticlePubMedPubMed CentralGoogle Scholar
  12. Ewing R, Cervero R. Travel and the Built Environment: A meta Analysis. J Am Plann Assoc. 2010;76(3):265–94.View ArticleGoogle Scholar
  13. Sugiyama T, Neuhaus M, Cole R, Giles-Corti B, Owen N. Destination and route attributes associated with adults' walking: a review. Med Sci Sports Exerc. 2012;44(7):1275–86.View ArticlePubMedGoogle Scholar
  14. Forsyth A, Hearst M, Oakes JM, Schmitz KH. Design and destinations: factors influencing walking and total physical activity.(Report). Urban Studies. 2008;45(9):1973.View ArticleGoogle Scholar
  15. Saelens BE, Handy SL. Built Environment Correlates of Walking: A Review. Medicine & Science in Sports & Exercise. 2008;40:S550–66.View ArticleGoogle Scholar
  16. Koohsari MJ, Badland H, Giles-Corti B. Viewpoint: (Re)Designing the built environment to support physical activity: Bringing public health back into urban design and planning. Cities. 2013;35:294–8.View ArticleGoogle Scholar
  17. Durand CP, Andalib M, Dunton GF, Wolch J, Pentz MA. Systematic review of built environment factors related to physical activity and obesity risk: implications for smart growth urban planning. Obes Rev. 2011;12(501):e173–82.View ArticlePubMedPubMed CentralGoogle Scholar
  18. Giles-Corti B, Vernez-Moudon A, Reis R, Turrell G, Dannenberg AL, Badland H, Foster S, Lowe M, Sallis JF, Stevenson M, et al. City planning and population health: a global challenge. Lancet. 2016;388(10062):2912–24.View ArticlePubMedGoogle Scholar
  19. Sallis JF, Bull F, Burdett R, Frank LD, Griffiths P, Giles-Corti B, Stevenson M. Use of science to guide city planning policy and practice: how to achieve healthy and sustainable future cities. Lancet. 2016;388(10062):2936–47.View ArticlePubMedGoogle Scholar
  20. Stevenson M, Thompson J, de Sá TH, Ewing R, Mohan D, McClure R, Roberts I, Tiwari G, Giles-Corti B, Sun X, et al. Land use, transport, and population health: estimating the health benefits of compact cities. Lancet. 2016;388(10062):2925–35.View ArticlePubMedPubMed CentralGoogle Scholar
  21. Forsyth ANN, Krizek KJ. Promoting Walking and Bicycling: Assessing the Evidence to Assist Planners. Built Environ. 2010;36(4):429.View ArticleGoogle Scholar
  22. Leal C, Chaix B: influence of geographic life environments on cardiometabolic risk factors: a systematic review, a methodological assessment and a research agenda. influence of geographic life environments on cardiometabolic risk factors: a systematic review, a methodological assessment and a research agenda 2011, 12(3):217-230.Google Scholar
  23. Hooper P, Knuiman M, Foster S, Giles-Corti B. The building blocks of a ‘Liveable Neighbourhood’: Identifying the key performance indicators for walking of an operational planning policy in Perth, Western Australia. Health and Place. 2015;36:173–83.View ArticlePubMedGoogle Scholar
  24. McCormack G, Giles-Corti B, Bulsara M. The relationship between destination proximity, destination mix and physical activity behaviors. Prev Med. 2008;46:33–40.View ArticlePubMedGoogle Scholar
  25. Cerin E, Leslie E, Ld T, Owen N, Frank LD. Destinations that matter: Associations with walking for transport. Health and Place. 2007;13(3):713–24.View ArticlePubMedGoogle Scholar
  26. Heath GW, Brownson RC, Kruger J, Miles R, Powell KE, Ramsey LT. The Effectiveness of Urban Design and Land Use and Transport Policies and Practices to Increase Physical Activity: A Systematic Review. J Phys Act Health. 2006;3:S55–76.View ArticlePubMedGoogle Scholar
  27. Giles-Corti B, Bull F, Knuiman M, McCormack G, Van Niel K, Timperio A, Christian H, Foster S, Divitini M, Middleton N, et al. The influence of urban design on neighbourhood walking following residential relocation: Longitudinal results from the RESIDE study. Social Science & Medicine. 2013;77:20–30.View ArticleGoogle Scholar
  28. Hirsch JA, Moore KA, Evenson KR, Rodriguez DA, Roux AVD. Research Article: Walk Score® and Transit Score® and Walking in the Multi-Ethnic Study of Atherosclerosis. American Journal Of Preventive Medicine. 2013;45:158–66.View ArticlePubMedPubMed CentralGoogle Scholar
  29. Knuiman MW, Christian HE, Divitini ML, Foster SA, Bull FC, Badland HM, Giles-Corti B. A longitudinal analysis of the influence of the neighborhood built environment on walking for transportation: the RESIDE study. In.: Johns Hopkins University Press. 2014;453Google Scholar
  30. Transportation Research Board: Does the built environment influence physical activity? Examining the evidence. In. Washington, DC, TRB.; 2005.Google Scholar
  31. Millward H, Spinney J, Scott D. Active-transport walking behavior: destinations, durations, distances. J Transp Geogr. 2013;28:101–10.View ArticleGoogle Scholar
  32. City of Portland: The Portland Plan. In. Portland; 2012.Google Scholar
  33. State Government of Victoria: Plan Melbourne - Metropolitan Planning Strategy. In. Victoria: Department of Transport; 2014.Google Scholar
  34. Commonwealth of Australia: Smart Cities Plan. In. Australia: Department of Premier and Cabinet; 2016.Google Scholar
  35. Cervero R: Transit-Orientated Development and the Urban Fabric. In: Sustainable railway Futures: Issues and Challenges. edn. Edited by Comtois C, Loo BPY, Hesse MP, Knowles RP: Farnham : Taylor and Francis, 2016.; 2015.Google Scholar
  36. Kamruzzaman M, Baker D, Washington S, Turrell G. Advance transit oriented development typology: case study in Brisbane, Australia. J Transp Geogr. 2014;34:54–70.View ArticleGoogle Scholar
  37. McNally M, Kulkarni A. Assessment of Influence of Land Use-Transportation System on Travel Behavior. Trans Res Rec. 1997;1607(1):105.View ArticleGoogle Scholar
  38. Charreire H, Weber C, Chaix B, Salze P, Casey R, Banos A, Badariotti D, Kesse-Guyot E, Hercberg S, Simon C, et al. Identifying built environmental patterns using cluster analysis and GIS: Relationships with walking, cycling and body mass index in French adults. International Journal of Behavioral Nutrition & Physical Activity. 2012;9:59–69.View ArticleGoogle Scholar
  39. Riva M, Apparicio P, Gauvin L, Brodeur J-M. Establishing the soundness of administrative spatial units for operationalising the active living potential of residential environments: an exemplar for designing optimal zones. Int J Health Geogr. 2008;7:43–3.Google Scholar
  40. Adams MA, Sallis JF, Kerr J, Conway TL, Saelens BE, Frank LD, Norman GJ, Cain KL. Neighborhood environment profiles related to physical activity and weight status: A latent profile analysis. Prev Med. 2011;52:326–31.View ArticlePubMedPubMed CentralGoogle Scholar
  41. Akar G, Chen N, Gordon SI. Influence of neighborhood types on trip distances: Spatial error models for Central Ohio. International Journal of Sustainable Transportation. 2016;10(3):284–93.View ArticleGoogle Scholar
  42. Growth Areas Authority: Precinct Structure Planning Guidelines: Preparing the Precinct Structure Plan. In. Victoria; 2009.Google Scholar
  43. Hooper P, Giles-Corti B, Knuiman M. Evaluating the Implementation and Active Living Impacts of a State Government Planning Policy Designed to Create Walkable Neighborhoods in Perth, Western Australia. Am J Health Promot. 2014;28:S5–S18.View ArticlePubMedGoogle Scholar
  44. Badland H, Mavoa S, Boulangé C, Eagleson S, Gunn L, Stewart J, David S, Giles-Corti B. Identifying, creating, and testing urban planning measures for transport walking: Findings from the Australian national liveability study. Journal of Transport &amp; Health. 2016;Google Scholar
  45. Cervero R, Kockelman K. Travel demand and the 3Ds: Density, diversity, and design. Transportation Research Part D. 1997;2:199–219.View ArticleGoogle Scholar
  46. The Urban Transport Institute and I-view Pty Ltd: Victorian Integrated Survey of Travel & Activity 2009-10: Survey procedures and documentation. In. Melbourne; 2011.Google Scholar
  47. Wei YD, Xiao WY, Wen M, Wei R. Walkability, Land Use and Physical Activity. Sustainability. 2016;8(1)Google Scholar
  48. Gunn LD, King TL, Mavoa S, Lamb K, Giles-Corti B, Kavanagh A. Identifying destination distances that support walking trips in local neighborhoods. J Trans Health. 2017;5:133–41.View ArticleGoogle Scholar
  49. Manaugh K, El-Geneidy A. Validating walkability indices: How do different households respond to the walkability of their neighborhood? Transportation Research Part D. 2011;16(4):309–15.View ArticleGoogle Scholar
  50. State Government of Victoria: Vicmap Property. In. Victoria: Department of Sustainability and Environment; 2012b.Google Scholar
  51. State Government of Victoria: Vicmap Transport. In. Victoria: Department of Sustainability and Environment; 2012c.Google Scholar
  52. Pitney Bowes Ltd: Axiom Business Points. In. Edited by Pitney Bowes Ltd. Sydney; 2014.Google Scholar
  53. Streiner D. Breaking up is hard to do: the heartbreak of dichotomizing continuous data. Canadian Journal of Psychiatry. 2002;47(3):262–6.View ArticlePubMedGoogle Scholar
  54. Ewing RH, Cervero R. Travel and the built environment: a meta-analysis. Journal of the American Planning Association. 2010;76(3):265–94.View ArticleGoogle Scholar
  55. Hair JF, Jr., Black WC, Babin BJ, Anderson RE: Multivariate data analysis: Harlow Pearson Education Limited, Seventh edition, Pearson new international edition.; 2014.Google Scholar
  56. Centers for Disease Control and Prevention: Physical activity and health: a report for the Surgeon General.. In. Atlanta: US Department of Health and Human Services; 1996.; 1996.Google Scholar
  57. Measuring Network Connectivity for Bicycling and Walking [http://reconnectingamerica.org/assets/Uploads/TRB2004-001550.pdf]
  58. Zhang M. On the Cul-de-Sac vs. Checker-Board Street Network: Search for Sustainable Urban Form. International review for spatial planning and sustainable development. 2013;1(1):1–16.View ArticleGoogle Scholar
  59. Messenger T, Ewing R. Transit-oriented development in the sunbelt. Trans Res Rec. 1996;1552:145–53.View ArticleGoogle Scholar
  60. Rodríguez DA, Aytur S, Forsyth A, Oakes JM, Clifton KJ. Relation of modifiable neighborhood attributes to walking. Prev Med. 2008;47:260–4.View ArticlePubMedGoogle Scholar
  61. Donovan S, Munro I. Impact of urban form on transport and economic outcomes . NZ Transport Agency. Research Report. 2013:513.Google Scholar
  62. Krizek K, Forsyth A, Baum L. Walking and Cycling International Literature Review: Final Report. In. Victoria: Department of. Transport. 2009;Google Scholar
  63. State Government of Victoria: Plan Melbourne - Metropolitan Planning Strategy. In. Victoria: Department of Transport; 2017.Google Scholar
  64. Buffel T, Phillipson C, Scharf T. Ageing in urban environments: Developing ‘age-friendly’ cities. Critical Social Policy. 2012;32(4):597–617.View ArticleGoogle Scholar
  65. Durand CP, Andalib M, Dunton GF, Wolch J, Pentz MA. A systematic review of built environment factors related to physical activity and obesity risk: implications for smart growth urban planning. Obesity reviews : an official journal of the International Association for the Study of Obesity. 2011;12(5):e173.View ArticleGoogle Scholar
  66. Macintyre S. Deprivation amplification revisited; or, is it always true that poorer places have poorer access to resources for healthy diets and physical activity? The international journal of behavioral nutrition and physical activity. 2007;4:32.View ArticlePubMedPubMed CentralGoogle Scholar
  67. Macintyre S, Ellaway A, Cummins S. Place effects on health: how can we conceptualise, operationalise and measure them? Social Science &amp. Medicine. 2002;55(1):125–39.Google Scholar
  68. Badland H, Turrell G, Giles-Corti B: Who does well where? Exploring how self-rated health differs across diverse people and neighborhoods. 2013.Google Scholar
  69. Europe WHOROf, World Health O: Closing the Gap in a Generation Health Equity through Action on the Social Determinants of Health. Geneva: World Health Organization; 2008.Google Scholar
  70. Ma L, Kent JL, Mulley C: Transport Disadvantage, Social Exclusion and Subjective Wellbeing: The Role of Built Environment - Evidence from Sydney, Australia. In: Australasian Transport Research Forum 2016 Proceedings: 16 – 18 November 2016 2016; Melbourne, Australia; 2016.Google Scholar
  71. King T, Bentley R, Thorton L, Kavanagh A. Does the presence and mix of destinations influence walking and physical activity? A multilevel study. International Journal of Behavioral Nutrition and Physical Activity. 2015;12:115.View ArticlePubMedPubMed CentralGoogle Scholar
  72. Cerin E, Cain K, Conway T, Van Dyck D, Hinckson E, Schipperijn J, De Bourdeaudhuij I, Owen N, Davey R, Hino A et al: Neighborhood environments and objectively measured physical activity in 11 countries. 2014, 46(12):2253-2264.Google Scholar
  73. King TL, Thornton LE, Bentley RJ, Kavanagh AM. The Use of Kernel Density Estimation to Examine Associations between Neighborhood Destination Intensity and Walking and Physical Activity. PLoS ONE. 2015;9Google Scholar
  74. Zenk SN, Schulz AJ, Matthews SA, Odoms-Young A, Wilbur J, Wegrzyn L, Gibbs K, Braunschweig C, Stokes C. Activity space environment and dietary and physical activity behaviors: A pilot study. Health and Place. 2011;17(5):1150–61.View ArticlePubMedPubMed CentralGoogle Scholar
  75. Hirsch JA, Winters M, Clarke P, McKay H. Generating GPS activity spaces that shed light upon the mobility habits of older adults: a descriptive analysis.(Research). Int J Health Geogr. 2014;13:51.View ArticlePubMedPubMed CentralGoogle Scholar
  76. Perchoux C, Chaix B, Cummins S, Kestens Y. Conceptualization and measurement of environmental exposure in epidemiology: accounting for activity space related to daily mobility. Health &amp; place. 2013;21:86.View ArticleGoogle Scholar
  77. Vallée J, Le Roux G, Chaix B, Kestens Y, Chauvin P. The ‘constant size neighbourhood trap’ in accessibility and health studies. Urban Studies. 2015;52(2):338–57.View ArticleGoogle Scholar
  78. Kwan M-P. The Uncertain Geographic Context Problem. Annals of the Association of American Geographers. 2012;102(5):958–68.View ArticleGoogle Scholar
  79. Meyer KA, Boone-Heinonen J, Duffey KJ, Rodriguez DA, Kiefe CI, Lewis CE, Gordon-Larsen P. Combined measure of neighborhood food and physical activity environments and weight-related outcomes: The CARDIA study. Health & Place. 2015;33:9–18.View ArticleGoogle Scholar

Copyright

© The Author(s). 2017

Advertisement