Master’s Thesis, 60 ECTS
Social-ecological Resilience for Sustainable Development Master’s programme 2014/16, 120 ECTS
Mapping neighbourhood typologies for social-ecological urbanism – A spatial experiential analysis of
Stockholm
Karl Samuelsson
Stockholm Resilience Centre
Research for Biosphere Stewardship and Innovation
Karl Samuelsson
Master’s thesis - Social-ecological resilience for sustainable development Stockholm Resilience Centre, Stockholm University
Supervisor: Stephan Barthel
Stockholm Resilience Centre Stockholm University
Co-supervisors: Ann Legeby
School of Architecture and the Built Environment Royal Institute of Technology
Lars Marcus
Department of Architecture
Chalmers University of Technology
Abstract. Studies on urban environments often display contradictory evidence regarding social and ecological outcomes, asserting conflicting development trajectories. In this thesis, affordance theory is applied with the aim of developing a method for relating high-precision mapping of urban structural characteristics to inhabitants’ experiences. I analyse
neighbourhood scale trade-offs and synergies between residential populations (RP), working populations (WP) and the ecosystem service temperature regulation (TR) in Stockholm municipality. Neighbourhood typology is introduced as an empirical classification of neighbourhoods based on these structural characteristics. I further analyse experiential outcome in different typologies by applying inhabitant experience data (N = 1828) from an online public participatory geographic information system survey. Analyses reveal strong trade-off patterns between populations and TR capacity. No typologies feature a large RP, a large WP and high TR capacity. Positive experiences are more likely in neighbourhoods with high TR capacity and negative experiences are more likely in neighbourhoods with a large WP, while most neighbourhoods are equally well experienced despite differences in services.
The thesis concludes that affordance theory provides methodological tools that when combined can close the gap between structural characteristics of the environment and experiential outcome, in turn leading to a better understanding of what constitutes social- ecological urbanism.
INTRODUCTION 4
Site 5
THEORY 7
The compact city paradox 7
The urban green area paradox 8
Linking people to the urban environment 9
METHODS 11
Space syntax 11
Public Participatory GIS 11
The neighbourhood and the place 11
Choosing variables 12
Study design 12
RESULTS 15
Relationships between services 15
Neighbourhood typologies 15
Inhabitants’ experiences of Stockholm 18
Experiences and individual services 18
Experiences and neighbourhood typologies 19
DISCUSSION 22
Implications for planning in Stockholm 24
Limitations 24
CONCLUSION 26
The case for affordance based analyses of urban environments 26
LITERATURE CITED 27
APPENDIX A – STUDY DESIGN 31
INTRODUCTION
Increased awareness of cities’ impact on the biosphere, both locally (Alberti 2005; Tratalos et al. 2007; Soga et al. 2014) and globally (Grimm et al. 2008; Seto et al. 2011; Seto et al. 2012;
Kennedy et al. 2015), as well as the quality of life for urban residents (van den Berg et al.
2007; Howley 2009; White et al. 2013; Alcock et al. 2014) presents a coupled challenge for urban planning: making cities more environmentally sustainable and more liveable. Also, urbanisation is projected to continue globally (United Nations 2014), requiring many cities to grow without compromising either environmental or social sustainability.
Current urban planning promotes high densities that decrease car dependency (Newman 2006), enable sustainable modes of transportation (Jabareen 2006) and require less energy- spending on heating (VandeWeghe and Kennedy 2007). Meanwhile, others argue for urban green areas’ importance for stress restoration purposes (van den Berg et al. 2007) and ecosystem service (ES) supply (Haase et al. 2014). This creates two objectives for research and urban planning alike. Firstly, an environment that potentially can provide environmental and social sustainability simultaneously is one that affords access for inhabitants to large populations and ES supplying areas (both of which I refer to as services), requiring spatially precise analyses. Secondly, while inhabitants’ experiences of the city relate to certain places, urban planning is often concerned with the neighbourhood scale, requiring experientially precise analyses that connect the two scales (Kyttä et al. 2013). This empirical study aims to present an analysis of Stockholm municipality that fulfils these objectives by mapping three kinds of data; 1) population distribution, 2) data on one ES, temperature regulation (TR) and 3) geocoded experiences of inhabitants. It answers these questions:
1) What are the trade-offs and synergies between populations and TR supplying areas on the neighbourhood scale in Stockholm municipality?
2) Can neighbourhoods be grouped into a set of typologies on the basis of these neighbourhood-scale services?
3) How do inhabitants’ experiences of places differ in relation to these services and typologies, respectively?
I hypothesise a general trade-off between populations and TR supplying areas on the
neighbourhood scale, with some neighbourhoods deviating from the general pattern. I further
hypothesise that neighbourhoods featuring a service mix will harbour more positive experiences than those with one predominant service.
Site
Stockholm municipality, Sweden, is a suitable study site, due to the mix of urban, suburban and nature environments, spatially precise available data on homes, workplaces and some ES, and the challenge to, considering a prospected population increase (Svensson et al. 2014),
“make development sustainable in the long term, from an economic, social and environmental perspective” (Stockholm City Planning Administration 2010:8).
At the end of 2014, 911 989 people lived in Stockholm municipality. It is divided into nine boroughs with population densities ranging from 2925 inhabitants per km² (Skarpnäck) to 15 842 (Södermalm) (City of Stockholm 2015). Nature areas are often interspersed with built-up
Figure 1. Map of Stockholm municipality. Land is white and water is grey. Important landscape features include the distinct division between inner city and suburbs, the Royal National City Park and Bromma airport.
areas, and large water bodies (grey in Figure 1) contribute to TR capacity. Important landscape features include the distinct division between inner city, southern and western suburbs, the Royal National City Park, and Bromma Airport (see Figure 1).
THEORY
The compact city paradox
Cities can be understood as emergent complex systems (Bettencourt and West 2010), or metaphorically as “organisms” (Batty 2012). This understanding has developed parallel to a critique of modernist top-down social engineering (Fainstein 2005). Jane Jacobs’ (1961) plea for a diverse cityscape regarding people and functions was seminal for this critique. Jacobs presented several criteria for achieving diversity but one, concentration of people, had a disproportionate impact on the urban planning paradigm of recent decades, the “compact city”
(Neuman 2005). Empirical inquiries regarding compact cities’ environmental sustainability has given mixed results (Neuman 2005). For example, Kennedy et al. (2015) found that lower electricity consumption per capita in compact cities was largely attributable to smaller gross floor areas per capita in such cities.
Evidence of social effects of population density is contradictory. Economic quantities such as information, innovation and wealth scale exponentially with city population (Bettencourt et al. 2007). However, while Bettencourt and West (2010) claim that city population growth induces denser settlements, a study involving 386 European cities found no such relationship (Fuller and Gaston 2009). Ample evidence exists of the beneficial effects of access to urban
Figure 2. The compact city can be seen as a paradox as it is often argued to be sustainable due to e.g. transport reasons while on the other hand lack of access to green space has been linked to e.g. detrimental health outcomes. Pictures courtesy of Pix Spotting at Flickr (left) and Timo at Flickr (right).
green areas on health, both physical (Mitchell and Popham 2008; Sugiyama et al. 2008) and mental (Alcock et al. 2014), as well as psychological restoration (Kaplan 1995; van den Berg et al. 2007). Thus, outcomes of compact cities appear paradoxical from a sustainability viewpoint – they are often less energy intensive but can reduce residents’ quality of life.
Contrarily, urban sprawl has been associated with increased social interaction (Brueckner and Largey 2008), and decreased physical activity (Ewing et al. 2003). Optimal centrality theory (Cicerchia 1999) attempts to look past the dichotomy between compact and sprawling cities by suggesting an urban density that strikes a balance between providing service access and reducing congestion, implying different social outcomes of urban densification in different contexts (McCrea and Walters 2012). Considering this ambiguous evidence, it seems safe to argue that sustainable urban development cannot be summarised in prescriptions for urban form alone (Jacobs 1961; Fainstein 2005; Neuman 2005).
The urban green area paradox
Ecosystem services (ES) are benefits people obtain from ecosystems (Millennium Ecosystem Assessment 2005). A growing body of literature highlights the importance of urban ES (Haase et al. 2014). While urban planning is increasingly giving consideration to cultural ES, such as recreation, regulating ES, such as TR or water mitigation, are seldom considered (Andersson et al. 2014a). Supply of these impact human well-being in cities (Pataki et al.
2011; Haase et al. 2014), varying with population density and urban form (Alberti 2005;
Tratalos et al. 2007). However, although well-proven indicators exist (Andersson et al.
2014b), few studies map these at high enough resolutions to be relevant for urban planning (Haase et al. 2014).
The framework of land-sparing contra land-sharing, originating in agricultural land-use research, has recently been applied to cities (Lin and Fuller 2013). It runs counterpart to the planning debate on compact contra sprawling urban development, from the perspective of biodiversity and ES (see Figure 2). Land-sparing strategies preserving large contiguous green areas are better for maintaining many regulating ES (Stott et al. 2015). However, inhabitants of land-sparing environments interact with nature less frequently (Soga et al. 2015). This second paradox boils down to a challenge of maintaining regulating ES supply whilst reaping social benefits of accessible green areas. Hence, regulating ES were considered for analysing trade-offs and synergies in neighbourhoods.
Linking people to the urban environment
For addressing these paradoxes, theoretical frameworks must incorporate the spatial and experiential dimensions of the human-environment relation. While emergent urban properties can be described as generalizable laws (Batty 2012; Bettencourt 2013), such approaches fall short as they are detached from local circumstances. Instead, affordance theory, originally conceived by Gibson (1979), is here applied as outlined by Chemero (2003) to study
structural characteristics and inhabitant experiences. Affordance theory emphasises contextual sensitivity, viewing human-environment relations as transactional and dynamic. Affordances are relations between abilities of organisms and features of environments (Chemero 2003). As such, they are phenomena anchored to unique situations. The concept embodies actions, emotions and states of mind that are influenced by bodily capacities, the environment’s physical structure and the socio-cultural context (Heft 2001).
Urban landscapes are mosaics of different land-uses (Andersson et al. 2014a), owing their qualities to a mix of social, ecological and constructed features. Even though the cultural ES concept is often applied to the urban context (Haase et al. 2014), the affordance concept arguably has greater possibilities to reconcile with this mix of features, surmounting the commonplace view of cities and nature as opposites. Moreover, urban inhabitants often act as
Figure 3. Examples of a compact city and land-sparing (left) and sprawl and land-sharing (right). The paradox of urban green areas is that in land-sparing environments, were biodiversity and ES are better maintained, people experience green areas less frequently. Pictures courtesy of Anthony Quintano at Flickr (left) and thisisbossi at Flickr (right).
stewards of social-ecological systems (Barthel et al. 2010), which is in line with affordance theory’s transactional view of the person-environment relation.
While the ES framework has until recently lacked a focus on spatial precision (Andersson et al. 2014b), much of urban design theory presents simplistic views on experiential outcomes of urban form, from modernism’s social engineering (Fainstein 2005) to the compact city
paradigm’s preoccupation with density (Neuman 2005). Urban environments planned with the assumption about the person-environment relation being unidirectional have consistently produced contradictory evidence. The transactional approach of affordance theory has
potential to unravel generalisable information from these contradictions without elevating it to panaceas for every context.
METHODS
Space syntax
Space syntax encapsulates theoretical and methodological discourses departing from the notion that human societies use space as a key resource for organizing themselves (Bafna 2003). Space syntax’ perceptual conceptualisation of space is corroborated by the person- environment relation of affordance theory (Marcus 2015). It has been widely applied in urban design research, showing that topological features of street networks are better predictors of movement than metric distances (Penn 2001). It has also been used to empirically test Jacob’s diversity criteria (Sayyar and Marcus 2013). This study applies space syntax methodology for investigating what populations neighbourhoods afford access to; rather than calculating from a point as the crow flies, it is measured by following streets or paths from that point.
Public Participatory GIS
Affordance theory is applied to investigate perceived environmental qualities, including qualities generated by ecosystems, the built environment, and a mix of both. Thus, as studied here these include cultural ES. Several studies have mapped urban environmental qualities through public participatory geographic information systems (PPGIS) (see Brown et al., (2014) for an overview). The principal aim of this methodology is to make GIS methods convenient to use locally to increase participation and inform land-use decisions (Brown and Kyttä 2014). However, it has managed to produce large datasets of experiential knowledge (Kyttä et al. 2013), and can thus provide more general insights for how different environments are experienced.
The neighbourhood and the place
Neighbourhood and place are concepts without formal definitions. Here, I define
neighbourhood to be everything within 500 metres distance from a point, according to what Gehl (2010) calls an acceptable walking distance and identifies as the approximate size of many city centres. I define place to be everything within 50 metres distance from a point, in line with some recent urban PPGIS studies (Broberg et al. 2013; Kyttä et al. 2013).
Appropriate methods for calculating distance depends on the measured variable; for populations the space syntax method was used while for TR, I measured as the crow flies.
This way of measuring services by mixing conventional and space syntax methods is
assessing the neighbourhood’s potential to provide affordances for its inhabitants. With a mix of services present on the neighbourhood scale, more desirable affordances should reasonably get realised. Hence the hypothesis previously stated, that more positive experiences should be found in neighbourhoods with a service mix.
Choosing variables
Jacobs (1961) argued for several benefits of mixing a residential population (RP) and a working population (WP); safer streets, more economic activity and leisure opportunities.
Mixed use has since become a fundamental concept within urban planning. Although its status as a panacea has been criticised (Wessel 2009), its prevalence needs to be
acknowledged. Hence, RP and WP were considered separate services, and I assessed accessibility to them separately.
The selection criteria for suitable regulating ES were 1) documented accurate indicators, 2) availability of high quality spatial data, 3) relevance for urban planning and 4) plausibility of impacting the experience of the city. Temperature regulation (TR), meaning the potential that ecosystems have to regulate local temperatures (Elmqvist and Mcdonald 2013), was the only ES deemed to fit well enough. A study on Stockholm showed persistently high temperatures to increase mortality rates (Rocklöv et al. 2011), hence TR likely being increasingly relevant for urban planning in Stockholm as it can aid adaptation to a future warmer climate.
Study design
A detailed account of the study design can be found in Appendix A.
Figure 4 presents an overview illustration of the study design.
I used a GIS-layer with RP and WP data for all addresses within Stockholm municipality, and another GIS-layer with TR values for all surfaces within Stockholm municipality (Table 1). A grid was superimposed on the land area of Stockholm municipality, dividing it into 10x10 metre squares. From the centre of each square, I measured accessible RP and WP and the surrounding neighbourhood’s TR capacity. Thus, each square obtained three values, together forming a profile of the surrounding neighbourhood. TR capacity was measured on a scale
from 1 (“no or meagre”) to 5 (“very large”), borrowed from Barthel et al. (2015b). As full TR values was not captured for squares with centres closer than 500 metre to the municipality border, these were excluded from the analysis.
Services were compared pair-wise to identify trade-offs and synergies. The data was
normalized and cluster analysis performed to identify neighbourhood typologies with similar characteristics. I subjectively named these to connote the character of much (although not all) of the landscapes composing them. The mean, minimum and maximum values for each variable in each typology were obtained to understand how the typologies relate to the services.
I utilised data from a PPGIS survey to map experiences of places within the study area. The survey was featured at Färgfabriken art hall during the autumn of 2015 and accessible online during and after the exhibition. It was designed to capture affordances. In the survey these
Table 1. Variables analysed as neighbourhood-scale services
Figure 4. An overview of the study design. The boxes at the top are sources of the data. The coloured fields represent the different kinds of spatial data applied or developed. Arrows and lines connecting the fields describe the kinds of analyses performed.
were called positive and negative experiences that people living in Stockholm have of places they are regularly exposed to (see Figure 5). I also utilised data on how three different factors – 1) buildings, streets and squares; 2) nature and 3) social interactions - influenced
experiences. This information was given by respondents on a scale from 1 (none at all) to 10 (very much).
I analysed relationships between neighbourhood services and inhabitants’ experiences.
Neighbourhood values for RP, WP and TR, as well as typology, were obtained for the location of each experience. In addition, RP, WP and TR values of the surrounding place of each experience were calculated, with the same methods as neighbourhood values. This allowed for a comparison of place- and neighbourhood-scale services’ respective relationship to experiences. Logistic regression was used to statistically explore the distribution of positive and negative experiences in relation to services and typologies. Ordinal regression was used to explore relationships between self-reported influencing factors of experiences and
typologies. These analyses controlled for age group and gender.
Figure 5. A screenshot from the online PPGIS survey. After the respondent marks a place where an experience occurs he or she may describe attributes of the experience.
RESULTS
Relationships between services
Hexagonal binning plots of pairwise relationships between services reveal logarithmic trade- off patterns between TR and populations (Figure 6). Few neighbourhoods combine
substantial RP or WP with “considerable TR capacity” (TR=3). However, TR coexist more with RP than WP. For TR values between 3 and 1.5, RP is generally an order of magnitude greater than WP (e.g. TR = 2 generally corresponds to RP around 2000 and WP around 200).
Also, the maximum WP value (44,215) is several times greater than the maximum RP value (17,058), indicating an extreme concentration of WP in the landscape.
Although the relationship between RP and WP is overall positive, trade-offs primarily influence the dataset, as only 0.32% of the study area have higher values for all services than the dataset centroid (RP = 1618, WP = 1048, TR = 2.195) (i.e. featuring overall synergy), compared with 20.0% having lower values for all services (Figure 7). Most areas with high values are found in the transition zone between city and suburbs, while areas with low values are generally found in suburban areas.
Neighbourhood typologies
From the cluster analysis, 7 typologies were identified (Figure 8a). Areas covered by typologies differ between 1.00% and 33.9% of the study area. Figure 8b shows the range of different services within typologies with darker coloured lines representing the typology average.
Counts
3 2,5 2 1,5 1
10 000 1000 100 10 1 3,5
Average temperature regulation
Working population (log scale) Counts
10 000 1000 100 10
1
10 000 1000 100 10 1
Working population (log scale)
Living population (log scale)
Counts
3 2,5 2 1,5 1
10 000 1000 100 10 1 3,5
Average temperature regulation
Living population (log scale)
Figure 6. Hexagonal binning plots of the pairwise relationships between neighbourhood services. RP and WP numbers are presented on base-10 logarithmic scales as reachable population within 500 metres from each measurement point. TR is presented on the same scale as the one used in Barthel et al. (2015a); 1 = no or meager TR capacity, 2 = some TR capacity, 3 = considerable TR capacity, 4 = large TR capacity, 5 = very large TR capacity.
Nature retreats features virtually no RP or WP but high TR values. Sprawling suburbs features substantially lower TR values together with slightly higher RP. Mixed suburbs follows this trend, but both these typologies have almost no WP. Urban pockets features a service mix without any extreme values. Jacob’s city features the kind of environment Jacobs (1961) advocated – a mix of RP and WP – whereas Downtown has extreme values for WP and low values for other services. These six typologies make up a gradient where TR is traded off for RP and WP. The seventh typology, Brownfields, falls outside this gradient with solely low values. No typology consistently features high values for all services, but areas with above-average values for all services are mostly found in Mixed suburbs. Areas with below- average values for all services are mostly found in Brownfields and Mixed suburbs.
Higher values than average for all three services Lower values than average for all three services Border of study area
Figure 7. The neighbourhood scale service values for the centroid of the dataset are TR = 2.195, RP = 1618, WP = 1048.
Green areas (0.32% of study area) have higher values for all three services, while red areas (20.0% of study area) have lower values for all three services. Green areas mostly overlap with the Mixed suburbs typology, while red areas mostly overlap with the Brownfields and Mixed suburbs typologies.
Neighbourhood typologies
Nature retreats Sprawling suburbs Mixed suburbs Brownfields Urban pockets Jacob’s city Downtown
Figure 8a (top). Neighbourhood typologies based on cluster analysis of neighbourhood scale services. Their areal extent as a proportion of the study area is: Nature retreats = 9.8%, Sprawling suburbs = 25.6%, Mixed suburbs = 33.9%, Brownfields = 12.7%, Urban pockets = 11.9%, Jacob’s city = 4.83%, Downtown = 1.00%.
Figure 8b (bottom). Normalised neighbourhood scale service values for each typology, where 1 represents maximum and 0 minimum. The coloured fields represent the upper and lower limit within each typology, while the darker coloured lines show the average of each typology.
Inhabitants’ experiences of Stockholm
1,043 respondents recorded 1,828 place-specific experiences within the study area (Figure 10). 71.9% (1314) are positive and 28.1% (514) are negative. Experience density is generally far higher in central areas than suburbs. While positive experiences are scattered, negative experiences are often concentrated around transportation hubs or along main travel routes.
Experiences and individual services
Out of all respondents, 452 also provided gender and age group data, together recording 1,032 experiences. This subset was used for logistic regression analysis (Table 2). Men and women were almost equally represented. However, people aged 25 – 44 were disproportionate, constituting 57% of the sample. Each service was analysed separately. For RP and WP, a base-2 logarithmic scale was used to facilitate interpretation, while for TR, the scale used earlier was kept.
Positive experience Negative experience Border of study area
Figure 10. 1043 respondents recorded 1828 place specific experiences within the study area. 71.9% (1314) were positive and 28.1% (514) were negative.
For neighbourhood services, the analysis revealed that the probability of an experience being positive increases with TR (OR = 5.99; p < 0.001), but decreases with RP (OR = 0.935; p <
0.01) and WP (OR = 0.717; p < 0.001). For the same services on the place-scale, the relationships are similar but less pronounced for TR (OR = 3.14; p < 0.001) and WP (OR = 0.868; p < 0.001). However, for RP, the relationship is reversed compared to the
neighbourhood scale (OR = 1.11; p < 0.001).
Experiences and neighbourhood typologies
Both amount and proportion of positive and negative experiences within each typology varied markedly (Table 3). Urban pockets contained the highest total number of experiences (429) while Nature retreats contained fewest (84). Nature retreats had the highest proportion of positive experiences (96.4%), while Downtown had the highest proportion of negative experiences (67.1%). Experience density ranged from 5.71 experiences per km² (Nature retreats) to 140 experiences per km² (Downtown).
Logistic regression (respondents = 452, N=1032) was carried out (Table 4) with Mixed suburbs as baseline, as this typology is the most prevalent and closest to
Table 2. Logistic regression models of the probability of experiences being positive or negative in relation to
neighbourhood and place scale services. Each line represents a different model. For RP and WP, base 2 logarithmic scales were used; a doubling of population change the odds of an experience being positive by the odds ratio. For TR, a five step scale was used (see Barthel et al. 2015a); for each step higher on the scale, the odds of an experience being positive changes by the odds ratio.
Table 3. Amount and proportion of positive and negative experiences as well as experience density, varied markedly across different neighbourhood typologies.
the dataset average. The analysis revealed that the probability of an experience being positive is significantly higher in Nature retreats (OR = 11.6; p < 0.001) and Sprawling suburbs (OR
= 5.82; p < 0.001), and significantly lower in Downtown (OR = 0.131; p < 0.001) and Brownfields (OR = 0.421; p < 0.001). For Urban pockets and Jacob’s city, there was no significant difference.
543 respondents recorded scores from 1 to 10 for environmental factors influencing 885 experiences. These varied substantially between typologies. Figure 12 shows the aggregated scores of each environmental factor in the typologies. In Nature retreats, Sprawling suburbs
Table 4. Logistic regression model of the probability of an experience being positive in different neighbourhood typologies, with Mixed suburbs being the baseline typology. Respondents = 452, N = 1032. *** p < .001
Figure 12. 543 respondents recorded scores for influencing environmental factors for 885 experiences. This chart shows all positive and negative experiences, weighted by influencing environmental factors, for different neighbourhood typologies. Each experience gets a weight between 1 and 10 for each influencing environmental factor that corresponds with the score given by the respondent for that experience. Dark blue represents buildings, streets and squares, green represents nature and red represents social interactions.
and Mixed suburbs, nature is the most important contributing factor for positive experiences.
In Urban pockets, all three factors contribute almost equally, while in Jacob’s city, Downtown and Brownfields, buildings, streets and squares as well as social interactions contribute more.
For negative experiences, buildings, streets and squares contribute most in Mixed suburbs, Urban pockets and Brownfields, while social interactions contribute slightly more in Jacob’s city and Downtown.
Ordinal regression, with Mixed suburbs as baseline, revealed significant differences in how environmental factors contribute to experiences in different typologies (Table 5). The number of experiences with a contributing environmental factor varied from 237 (negative
experiences and social interactions) to 518 (positive experiences and nature).
Gradients exist in the likelihood of environmental factors contributing to positive experiences, largely reflecting services, with buildings, streets and squares and social interactions
contributing more in densely populated neighbourhoods while nature contributes more in neighbourhoods with high TR. Far fewer differences are significant for factors contributing to negative experiences. In Nature retreats and Jacob’s city, negative experiences are less likely due to buildings, streets and squares, while in Sprawling suburbs, Jacob’s city and
Downtown, they are more likely due to social interactions.
Table 5. Ordinal regression models of the probability of different environmental factors contributing to experiences in different neighbourhood typologies, with Mixed suburbs being the baseline typology. Separate models were used for each environmental factor, and for positive and negative experiences, respectively, giving a total of six models. *p<.05,
**p<.01, ***p<.001
DISCUSSION
The analyses revealed strong trade-off patterns between populations and TR. However, trade- offs between WP and TR is more extreme than between RP and TR. This certainly does not reflect greater difficulties of spatially combining WP and TR in neighbourhoods, but rather an extreme concentration of WP resulting from planning decisions (Gullberg 2002). The cluster analysis revealed exceptions to the trade-off pattern to not be common enough for any cluster to feature consistently high service levels. Indeed, less than 1% of the study area had above- average values for all services, showing that no epitome typology for reaping benefits of both compact neighbourhoods and TR exists within the study area. This does not however imply that such a neighbourhood would be impossible to construct, as smaller floor areas per capita potentially could reduce energy consumption for heating, leave space for ample vegetation and provide high enough densities for sustainable transportation.
On the neighbourhood scale, higher RP and WP values increase the likelihood of experiences being negative, with WP having more adverse effects than RP. This might be because of either of two explanations, or a combination of them. Firstly, that WP is highly concentrated, hence neighbourhoods with large WP being very dense and having limited possibilities for stress restoration. This is in line with several environmental psychology studies (Kaplan 1995;
van den Berg et al. 2007; Nordh et al. 2009). Secondly, that neighbourhoods with large WP lack the neighbourhood-scale mix of uses that Jacobs (1961) advocates. Whether one or both of these explanations account for the observed outcome, planning strategies to mitigate negative experiences ought to be similar in not striving for extreme concentration of WP in the landscape. Conversely, an opposite strategy, where services are mixed, potentially produces neighbourhoods that are vibrant outside work hours, yet provide possibilities for stress restoration.
On the place scale, large RP increase the likelihood of an experience being positive,
contrasting the neighbourhood scale relationship. The explanations outlined above can also be applied here. Place-scale RP might provide the social street presence that Jacobs (1961) argue is crucial for feeling secure, whereas a high neighbourhood-scale RP might limit possibilities for stress restoration. This seems to support the balance between density and congestion of optimal centrality theory.
Evidence from this study indicates that density alone is a poor indicator of experiential outcome in cities, as has been suggested elsewhere (Howley et al. 2009; Miles and Song 2009). The odds of an experience being positive is 7.63 times greater in Jacob’s city than Downtown, a typology of similar density (see Figure 13 for an example). The analysis of environmental factors’ impact on experiences showed that the factor with the largest difference in experiential outcome comparing Jacob’s city with Downtown is social
interactions. I hypothesise that the largely negative impact of social interactions in Downtown is due to stress caused by large amounts of people traveling to and from work simultaneously.
This could not be verified with data included in this study. However, density must clearly be accompanied by other structural measurements of urban form for rewardingly describing a neighbourhood.
The logarithmic trade-off pattern between TR and populations supports the argument that land-sparing is important for maintaining regulating ES (Stott et al. 2015), as even modest populations sharply decrease a neighbourhood’s TR capacity. This study seems to also support the finding of Soga et al. (2015), that people experience nature more frequently in land-sharing areas, as nature generally contributes more to meaningful experiences in typologies closer to land-sharing (Sprawling suburbs and Mixed suburbs) than it does in typologies closer to land-sparing (Nature retreats, Urban pockets, Jacob’s city and
Downtown). However, the evidence is not completely unanimous, as nature contributes more in Urban pockets than in Mixed suburbs, possibly due to higher experience densities in green areas within this typology.
Figure 13. The odds of an experience being positive is 7.63 times greater in Jacob’s city compared to Downtown, a typology of similar density. The left picture shows a place with a recorded experience just south of Stockholm Central station, in Downtown, where RP is 282 persons, WP is 15 378 persons and total population within 500 metres is 15 660 persons. The right picture shows Surbrunnsgatan from Norrtullsgatan, a place in Jacob’s city with a recorded experience, where RP is 9618 persons, WP is 5970 persons and total population within 500 metres is 15 588 persons. The pictures where taken around 11 a.m. April 29th 2016.
Implications for planning in Stockholm
Stockholm municipality aims to build 140,000 new housing units until 2030 (Svensson et al.
2014), an increase of 30.4% compared to the 2014 stock (City of Stockholm 2016). Different strategies for achieving this increase will have vastly different impacts on both TR supply and inhabitants’ experiences. Brownfields constitutes 12.7% of the study area. Development projects within this typology might increase RP, WP and TR simultaneously while also benefitting experiential outcome. These findings justify Stockholm municipality’s larger urban planning projects in recent years that have focused on converting former harbour and industrial areas into mixed-use neighbourhoods (e.g. Norra Djurgårdsstaden) (Stockholm City Planning Administration 2010). Although other services not investigated in this study might influence the cogency of this strategy, it is certainly favourable when compared to the possible loss of important TR capacity and positive experiences by developing large contiguous green areas.
Apart from conversion of former industrial areas, Stockholm municipality focuses on gradual densification of suburban areas through infill projects (Stockholm City Planning
Administration 2010). As there are no existing neighbourhoods combining high levels of RP, WP and TR, densification of these neighbourhoods will probably have negative impacts on TR capacity. However, depending on level of densification and post-development service mix, it might avoid negative impacts on inhabitants’ experiences. The spatially and
experientially precise analysis methods introduced here can potentially aid these infill projects in addressing service trade-offs as well as local experiential knowledge.
Lastly, Stockholm’s city centre, overlapping with the Downtown typology, is currently undergoing further densification (Stockholm City Planning Administration 2010). These developments are mostly office and retail projects, attracting criticism from among others Stockholm Beauty Council (Tottmar 2014). Considering the evidence presented here, such developments seem highly questionable as they risk exacerbate experiential outcomes in central Stockholm.
Limitations
The different kinds of sourced data were produced in different years. It is reasonable to believe that the data for RP (from 2009) and WP (from 2008) are somewhat inconsistent with
today’s numbers in former brownfields areas that have undergone or are undergoing urban renewal.
Only one ES was investigated and future studies might broaden the palette of ES, as well as including cultural amenities, welfare amenities (similar to Kytta et al., 2015) or socio- economic data (similar to Miles et al., 2009), to broaden and deepen the understanding of trade-offs and synergies in the urban landscape.
As stated, this study seems to support the notion that people experience nature more
frequently in land-sharing areas. However, a methodology of even greater spatial refinement is needed for better placement of typologies on a land-sharing-land-sparing gradient. This methodology would account for whether services are evenly distributed or concentrated to a small area within a neighbourhood. Also, the 500 metre definition of neighbourhood size might be challenged to explore if the land-sparing-land-sharing paradox disappears at different scales.
Respondents were disproportionately aged 25 – 44, while far fewer were aged under 18 or over 70. There is no way to verify if the sample was socio-economically representative of Stockholm. As the survey was publicly accessible, it is also not possible to exclude ideological bias resulting from interest group mobilisation, something that has previously manifested in PPGIS studies (Brown et al. 2013).
CONCLUSION
The case for affordance based analyses of urban environments
This study shows a strong coupling of structural characteristics of urban environments with inhabitants’ experiences. It also shows that experiential outcome is not easily captured by one-dimensional structural measurements, and is highly context-dependent, as stipulated in affordance theory. By combining two methodologies springing from affordance theory - space syntax and PPGIS - in a novel way, typologies with different potentials for realising
affordances could be spatially mapped and corroborated by realised affordances. The insights uncovered by logistic regression analysis of typologies and experiences illustrates this
method’s usefulness. While positive experiences are more likely in Nature retreats and Sprawling suburbs and negative more likely in Downtown and Brownfields, differences between Mixed suburbs, Urban pockets and Jacob’s city are not statistically significant. Thus, most of the study area is equally well experienced despite large variances in services,
illustrating that it is impossible to make predictions about experiential outcome of service changes without considering the context. This complex relationship would have been overlooked without the clustering of neighbourhoods into typologies.
By viewing Jacob’s writings through an affordance theory lens, it is possible to go beyond oversimplified interpretations that lead to the compact city paradigm. Moreover, this lens resolves the inadequate division between city and nature by considering these as two components of a motley environment. Lastly, applying the space syntax method for
measuring accessibilities provides a way to study PPGIS-generated affordance data according to the notion of a perceptual space from affordance theory. The mapped experiences become snapshots of a dynamic environment rather than isolated phenomena in an isotropic space, closing the gap between structural characteristics of the environment and experiential outcome.
LITERATURE CITED
Alberti, M. 2005. The Effects of Urban Patterns on Ecosystem Function. International Regional Science Review 28: 168–192.
Alcock, I., M. P. White, B. W. Wheeler, L. E. Fleming, and M. H. Depledge. 2014.
Longitudinal Effects on Mental Health of Moving to Greener and Less Green Urban Areas. Environmental Science & Technology 48: 1247–1255.
Andersson, E., S. Barthel, S. Borgström, J. Colding, T. Elmqvist, C. Folke, and Å. Gren.
2014a. Reconnecting cities to the biosphere: stewardship of green infrastructure and urban ecosystem services. Ambio 43: 445–53.
Andersson, E., T. McPhearson, P. Kremer, E. Gomez-Baggethun, D. Haase, M. Tuvendal, and D. Wurster. 2014b. Scale and context dependence of ecosystem service providing units. Ecosystem Services 12: 1–8.
Bafna, S. 2003. Space Syntax: A Brief Introduction to Its Logic and Analytical Techniques.
Environment & Behavior 35: 17–29.
Barthel, S., C. Folke, and J. Colding. 2010. Social-ecological memory in urban gardens- Retaining the capacity for management of ecosystem services. Global Environmental Change 20: 255–265.
Barthel, S., A. Koffman, M. Bovin, E. Lundqvist, E. Campbell, and M. Tuvendal. 2015a.
Appendix A . Teknisk bilaga - Kartläggning av ekosystemtjänster i Stockholms stad.
Stockholm.
Barthel, S., A. Koffman, M. Bovin, E. Lundqvist, E. Campbell, and M. Tuvendal. 2015b.
Kartläggning och analys av ekosystemtjänster i Stockholms stad. Stockholm.
Batty, M. 2012. Building a science of cities. Cities 29: 9–16.
van den Berg, A. E., T. Hartig, and H. Staats. 2007. Preference for nature in urbanized societies: Stress, Restoration, and the Pursuit of Sustainability. Journal of Social Issues 63: 79–96.
Bettencourt, L., and G. West. 2010. A unified theory of urban living. Nature 467: 912–913.
Bettencourt, L. M. A. 2013. The Origins of Scaling in Cities. Science 340: 1438–1441.
Bettencourt, L. M. A., J. Lobo, D. Helbing, C. Kühnert, and G. B. West. 2007. Growth, innovation, scaling, and the pace of life in cities. Proceedings of the National Academy of Sciences of the United States of America 104: 7301–7306.
Bonaiuto, M., F. Fornara, and M. Bonnes. 2003. Indexes of perceived residential environment quality and neighbourhood attachment in urban environments: A confirmation study on the city of Rome. Landscape and Urban Planning 65: 41–52.
Broberg, A., M. Kyttä, and N. Fagerholm. 2013. Child-friendly urban structures: Bullerby revisited. Journal of Environmental Psychology 35: 110–120.
Brown, G., and M. Kyttä. 2014. Key issues and research priorities for public participation GIS (PPGIS): A synthesis based on empirical research. Applied Geography 46: 126–136.
Brown, G., M. Kelly, and D. Whitall. 2013. Which “public”? Sampling effects in public participation GIS (PPGIS) and volunteered geographic information (VGI) systems for public lands management. Journal of Environmental Planning and Management 57:
190–214.
Brueckner, J. K., and A. G. Largey. 2008. Social interaction and urban sprawl. Journal of Urban Economics 64: 18–34.
Chemero, A. 2003. An Outline of a Theory of Affordances. Ecological Psychology 15: 181–
195.
Cicerchia, A. 1999. Measures of optimal centrality: Indicators of city effect and urban overloading. Social Indicators Research 46: 273–299.
City of Stockholm. 2015. Areal och befolkningstäthet i stadsdelsområden, SDN-delar och stadsdelar. <http://www.statistikomstockholm.se/images/stories/excel/b039.htm>
City of Stockholm. 2016. Statistisk årsbok för Stockholm - 2016. Stockholm.
Elmqvist, T., and R. I. Mcdonald. 2013. Urbanization, Biodiversity and Ecosystem Services:
Challenges and Opportunities. Edited by T. Elmqvist, M. Fragkias, J. Goodness, B.
Güneralp, P. J. Marcotullio, R. I. McDonald, S. Parnell, M. Schewenius, M. Sendstad, K.
C. Seto and C. Wilkinson. Dordrecht: Springer Netherlands.
Ewing, R., T. Schmid, R. Killingsworth, A. Zlot, and S. Raudenbush. 2003. Relationship between Urban Sprawl and Physical Activity, Sedentary Beahviour, and BMI. Obesity 15: 430–445.
Fainstein, S. S. 2005. Cities and Diversity: Should We Want It? Can We Plan For It? Urban Affairs Review 41: 3–19.
Fuller, R. A., and K. J. Gaston. 2009. The scaling of green space coverage in European cities.
Biology Letters 5: 352–355.
Gehl, J. 2010. Cities for people. Washington, D.C.: Island Press.
Gibson, J. J. 1979. The ecological approach to visual perception. Boston: Houghton Mifflin.
Grimm, N. B., S. H. Faeth, N. E. Golubiewski, C. L. Redman, J. Wu, X. Bai, and J. M.
Briggs. 2008. Global Change and the Ecology of Cities. Science 319: 756–760.
Gullberg, A. 2002. City - drömmen om ett nytt hjärta: moderniseringen av det centrala Stockholm 1951-1979. Stockholm: Stockholmia.
Haase, D., N. Larondelle, E. Andersson, M. Artmann, S. Borgström, J. Breuste, E. Gomez- Baggethun, Å. Gren, Z. Hamstead, R. Hansen, N. Kabisch, P. Kremer, J. Langemeyer, E.
L. Rall, T. McPhearson, S. Pauleit, S. Qureshi, N. Schwarz, A. Voigt, D. Wurster and T.
Elmqvist. 2014. A quantitative review of urban ecosystem service assessments:
Concepts, models, and implementation. Ambio 43: 413–433.
Heft, H. 2001. Ecological psychology in context: James Gibson, Roger Barker, and the legacy of William James’s radical empiricism. Mahwah: Lawrence Earlbaum Associates,
Publishers.
Howley, P. 2009. Attitudes towards compact city living: Towards a greater understanding of residential behaviour. Land Use Policy 26: 792–798.
Howley, P., M. Scott, and D. Redmond. 2009. Sustainability versus liveability : an examination of neighbourhood satisfaction. Journal of Environmental Planning and Management 52: 847–864.
Jabareen, Y. R. 2006. Sustainable Urban Forms: Their Typologies, Models, and Concepts.
Journal of Planning Education and Research 26: 38–52.
Jacobs, J. 1961. The Death and Life of Great American Cities. New York: Random House.
Kaplan, S. 1995. The restorative benefits of nature: Toward an integrative framework.
Journal of Environmental Psychology 15: 169–182.
Kennedy, C. A., I. Stewart, A. Facchini, I. Cersosimo, R. Mele, B. Chen, M. Uda, A. Kansal, A. Chiu, K. Kim, C. Dubeux, E. Lebre La Rovere, B. Cunha, S. Pincetl, J. Keirstead, S.
Barles, S. Pusaka, J. Gunawan, M. Adegbile, M. Nazariha, S, Hoque, P. J. Marcotullio, F. Gonzalez Otharan, T. Genena, N. Ibrahim, R, Farooqui, G, Cervantes and A. D. Sahin.
2015. Energy and material flows of megacities. Proceedings of the National Academy of Sciences of the United States of America 112: 5985–5990.
Kytta, M., A. Broberg, M. Haybatollahi, and K. Schmidt-Thome. 2015. Urban happiness:
context-sensitive study of the social sustainability of urban settings. Environment and Planning B: Planning and Design 47: 1–24.
Kyttä, M., A. Broberg, T. Tzoulas, and K. Snabb. 2013. Towards contextually sensitive urban densification: Location-based softGIS knowledge revealing perceived residential
environmental quality. Landscape and Urban Planning 113: 30–46.
Legeby, A. 2013. Patterns of co-presence: Spatial configuration and social segregation (Doctoral dissertation). Royal Institute of Technology.
Lin, B. B., and R. A. Fuller. 2013. Sharing or sparing? How should we grow the world’s cities? Journal of Applied Ecology 50: 1161–1168.
Marcus, L. 2015. Ecological space and cognitive geometry – Linking humans and environment in space syntax theory. In Proceedings of the 10th International Space Syntax Symposium, ed. K. Karimi, L. Vaughan, K. Sailer, G. Palaiologou, and T. Bolton.
London: Space Syntax Laboratory, The Bartlett School of Architecture, University College London.
McCrea, R., and P. Walters. 2012. Impacts of Urban Consolidation on Urban Liveability:
Comparing an Inner and Outer Suburb in Brisbane, Australia. Housing, Theory and Society 29: 190–206.
Miles, R., and Y. Song. 2009. “Good” neighborhoods in Portland, Oregon: Focus on both social and physical environments. Journal of Urban Affairs 31: 491–509.
Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well-being. Washington, DC.: Island Press.
Mitchell, R., and F. Popham. 2008. Effect of exposure to natural environment on health inequalities: an observational population study. Lancet 372: 1655–1660.
Neuman, M. 2005. The Compact City Fallacy. Journal of Planning Education and Research 25: 11–26.
Newman, P. 2006. The environmental impact of cities. Environment and Urbanization 18:
275–295.
Nordh, H., T. Hartig, C. M. Hagerhall, and G. Fry. 2009. Components of small urban parks that predict the possibility for restoration. Urban Forestry and Urban Greening 8: 225–
235.
Pataki, D. E., M. M. Carreiro, J. Cherrier, N. E. Grulke, V. Jennings, S. Pincetl, R. V. Pouyat, T. H. Whitlow and W. C. Zipperer. 2011. Coupling biogeochemical cycles in urban environments: Ecosystem services, green solutions, and misconceptions. Frontiers in Ecology and the Environment 9: 27–36.
Penn, A. 2001. Space Syntax and Spatial Cognition. In Proceedings. 3rd International Space Syntax Symposium, ed. J. Peponis, J. Wineman, and S. Bafna, 11.1–11.17. Atlanta: A.
Alfred Taubman College of Architecture and Urban Planning, University of Michigan.
Rocklöv, J., K. Ebi, and B. Forsberg. 2011. Mortality related to temperature and persistent extreme temperatures: a study of cause-specific and age-stratified mortality.
Occupational and Environmental Medicine 68: 531–536.
Sayyar, S. S., and L. Marcus. 2013. Designing difference : Interpreting and testing Jane Jacobs’ criteria for urban diversity in space syntax terms. In Proceedings of the Ninth International Space Syntax Symposium, ed. Y. O. Kim, H. T. Park, and K. W. Seo.
Seoul: Sejong University.
Seto, K. C., M. Fragkias, B. Güneralp, and M. K. Reilly. 2011. A Meta-Analysis of Global Urban Land Expansion. PLoS ONE 6: e23777.
Seto, K. C., A. Reenberg, C. G. Boone, M. Fragkias, D. Haase, T. Langanke, P. Marcotullio, D. K. Munroe, B. Olah and D. Simon. 2012. Urban land teleconnections and
sustainability. Proceedings of the National Academy of Sciences of the United States of America 109: 7687–7692.
Soga, M., Y. Yamaura, S. Koike, and K. J. Gaston. 2014. Land sharing vs. land sparing: does the compact city reconcile urban development and biodiversity conservation? Journal of Applied Ecology 51: 1378–1386.
Soga, M., Y. Yamaura, T. Aikoh, Y. Shoji, T. Kubo, and K. J. Gaston. 2015. Reducing the extinction of experience: Association between urban form and recreational use of public greenspace. Landscape and Urban Planning 143: 69–75.
Stockholm City Planning Administration. 2010. The walkable city: Stockholm City Plan.
Stockholm.
Stott, I., M. Soga, R. Inger, and K. J. Gaston. 2015. Land sparing is crucial for urban ecosystem services. Frontiers in Ecology and the Environment 13: 387–393.
Sugiyama, T., E. Leslie, B. Giles-Corti, and N. Owen. 2008. Associations of neighbourhood greenness with physical and mental health: do walking, social coherence and local social interaction explain the relationships? Journal of Epidemiology and Community Health 62: e9.
Svensson, N., B. Bergman, G. Jensen, M. Anheim, P. Krensler, M. Lundberg, P. Enarsson, G.
Landahl, A. Västberg, C. Carlander and A. Wojtczuk. 2014. Bostadspotential i Stockholm. Stockholm: City of Stockholm.
Tottmar, M. 2014. Skönhetsrådets attack: Bygget- liknar rovdrift. Dagens Nyheter, January 14. <http://www.dn.se/sthlm/skonhetsradets-attack-bygget-liknar-rovdrift/>
Tratalos, J., R. A. Fuller, P. H. Warren, R. G. Davies, and K. J. Gaston. 2007. Urban form, biodiversity potential and ecosystem services. Landscape and Urban Planning 83: 308–
317.
United Nations, Department of Economic and Social Affairs, Population Division. 2014.
World Urbanization Prospects: The 2014 Revision, Highlights (ST/ESA/SER.A/352).
VandeWeghe, J. R., and C. Kennedy. 2007. A Spatial analysis of residential greenhouse gas emissions in the toronto census metropolitan area. Journal of Industrial Ecology 11:
133–144.
Wessel, T. 2009. Does diversity in urban space enhance intergroup contact and tolerance?
Geografiska Annaler, Series B: Human Geography 91: 5–17.
White, M. P., I. Alcock, B. W. Wheeler, and M. H. Depledge. 2013. Would You Be Happier Living in a Greener Urban Area? A Fixed-Effects Analysis of Panel Data. Psychological Science 24: 920–928.
APPENDIX A – STUDY DESIGN
I used a GIS-layer with RP and WP data for each address within Stockholm municipality, created by Legeby (2013) from data sourced from Stockholm county council’s Growth and Regional Planning Administration (RP) and Stockholm municipality’s City Planning
Administration (WP). Numbers on RP were from 2009 and numbers on WP were from 2008.
In addition, I used a GIS-layer with streets and paths within Stockholm represented as lines, know in the space syntax literature as an axial map. This layer was created by the Spatial Analysis and Design group at KTH School of Architecture in 2012.
TR data was sourced from Stockholm municipality’s City Planning Administration. A map combining biotope data from 2009 with laser data on volume and proportional surface cover of vegetation from 2012 was used. The spatial units were squares of 2x2 metres. Each square had a multivariate score ranging from 1 to 5, consisting of a biotope score, a volume score and a surface cover score, each ranging from 1 to 5. The proportional weighting was 0,5 for biotopes and 0,25 for volume and surface cover respectively. For a more thorough description of methods used for producing this map, see Barthel et al. (2015).
A grid was superimposed on the land area of Stockholm municipality, dividing it into squares of 10x10 metres. From the centre of each square, I measured what addresses were reachable following streets and paths for 500 metres, and summed up RP and WP, respectively, for all reachable addresses. For this, I used the Place Syntax Tool plugin for Mapinfo version 10.
From the centre of each square, I made a circle with a 500 metre buffer distance, summed up all TR scores within it and divided by the area of the circle. This operation was made in QGIS version 2.12. Thus, each square obtained three values, together forming a profile of the
surrounding neighbourhood. All squares with centres closer than 500 metres from the municipality border were excluded from the analysis.
For the pair-wise comparison of services, I used the hexbin package in RStudio version 0.98 to produce hexagonal binning plots.
For the cluster analysis, I used the cluster package in RStudio. The data was transformed so that the minimum value of variables was set to zero and the maximum to one. K-means cluster analysis was performed with 5, 6, 7 and 8 clusters. The number of clusters to proceed
with was chosen after looking both at a scree plot and the visual outcome on a map, to correspond with low within group sum of squares and balance simplicity and complexity on the map.
A cluster analysis was performed with one cluster to know the centroid of the whole dataset.
With this reference point, percentages of data points with variable values above or below the dataset average was calculated.
The minimum, mean and maximum value of each cluster was used to illustrate the different profiles of the typologies. These diagrams were made in Microsoft Excel.
The PPGIS survey was designed and distributed by a project group at Stockholm Resilience Centre, of which I was part, to create a data pool for several studies. The survey was designed to record affordances as defined by Chemero (2003). Respondents were asked to insert data points representing affordances on a map. In the survey, these were called experiences.
Follow-up questions were designed to capture several characteristics of the experience. The respondent was asked to describe how three different factors (1. Buildings, streets and squares, 2. Nature and 3. Social interactions) influenced the experience, each on a scale from 1 (none at all) to 10 (very much). The respondent was also asked to input data not utilised in this study, including two additional groups of characteristics; those that describe the
respondents’ physical activities or psychological state and those that describe more detailed features of the environment. Further division into functional, social, aesthetic and atmospheric characteristics was done according to the Perceived Residential Environmental Quality scale from Bonaiuto et al. (2003), similar to a 2013 study in Helsinki (Kyttä et al. 2013).
Respondents were also asked to record age group, gender, nationality and time of living in Stockholm. There was no lower or upper limit for how many experiences respondents could record. There was no lower age limit for the survey. Experience characteristics and
demographic information was optional to insert.
The survey was publicly accessible online from September 21st 2015, and still is at the time of writing. It also featured at the exhibition Experiment Stockholm at Färgfabriken art hall in
southern Stockholm during the autumn of 2015. A Facebook page and Twitter account for the survey was created to spread awareness of its existence and continually disseminate input.
Several municipalities within Stockholm County helped raising awareness of the survey’s existence by spreading information online, on physical notice boards and in local newspapers.
While we acknowledge the uncertainty of the degree to which recorded data is representative of the whole population, our main objective was to maximize respondent numbers and demonstrate the possibilities of PPGIS in urban planning.
The survey dataset used for this study was downloaded on February 25th 2016.
Neighbourhood service values were added as attributes of each experience from the
intersecting grid square. This was done in QGIS. RP, WP and TR values of the surrounding place of each experience were calculated with the same methods as for neighbourhood values, but with the experiences as origin points instead of grid centroids.
Diagrams of experience distribution in different typologies were made in Microsoft Excel.
Logistic and ordinal regression was performed in RStudio. To facilitate interpretation, RP and WP was transformed to a base 2 logarithmic scale prior to analysis. TR values were kept on the five step scale used earlier. For ordinal regression analysis, I used the MASS package.
