Space Syntax  



  New findings on the link between crime and design challenge the widely accepted benefits of 'defensible space'. Space Syntax studies show that streets are made safer through 'natural surveillance' by neighbours and passers-by. Detailed studies of crime patterns are being used to identify problematic areas and understand the factors which influence safety.

....................................................... ............. .......RELATED PROJECTS FULL PROJECT PORTFOLIO

DOMESTIC BURGLARY ....................... ..................... ..................... ........................ analysis of crime location and spatial accessibility, UK



To facilitate an informed debate on the issue of crime and spatial design, Space Syntax has decided to create an online database of studies so that people can access and assess the first hand evidence for themselves. The first study on the site is "DO BURGLARS UNDERSTAND DEFENSIBLE SPACE?" by Bill Hillier and Simon Shu. This paper was originally published in the April 1999 edition of Planning In London and is reproduced with permission (contact 020 7834 9471 for information on Planning in London).

Additional studies will be added to the site as soon as they are available. The forthcoming article "CRIME AND URBAN LAYOUT: THE NEED FOR EVIDENCE" by Shu and Hillier, due to be published in the forthcoming book from the Institute for Public Policy Research (IPPR): 'Key Issues in Crime Prevention, Crime Reduction and Community Safety' (eds. Ballantyne S, MacLaren V, Pease K), will be soon available in PDF format.


Crime and space research has poor record for a number of reasons. First, too little effort has been made to isolate the effects of spatial design from those of social composition. In residential crime this is critical. Markets forces and local bureaucracies alike tend to ensure that better of people end up in good areas and less well off people in less good areas. Since more crime occurs in less good areas, it is hard to distinguish the effects of space from those of the social process that put people there. Second, too little attention has been given to spatial detail since studies have been at the level of the estate or the area rather than the individual crime. Third, too many different kinds of crime have been lumped together. It is likely that each type of crime has its own spatial logic which need to be identified. Fourth, studies have been too loose statistically in dealing with the interaction between factors, and so the work has been criticised.

The first paper on the database deals with the issue of domestic burglary (and to a lesser extent car crime) and how it is facilitated or inhibited by the design of public space. Are cul-de-sacs really safer than through roads? Does it help to separate vehicles from pedestrians? Do you really benefit from having small groups of neighbours, or are you safer with more ? Should we group houses into 'territorial' units? Or is it better to be on a more anonymous street?

There are four problems to be solved to show an influence from spatial layout on crime and begin to answer these questions. First, you have to take out the background effects of social composition and location. Burglary rates for inner urban areas are10.3%, 6.3 for 'suburban' areas and 3.9 for rural. Owner-occupiers (5.3%) with good incomes (5.6%) are much less likely to be burgled than renters (8.9%) on low incomes (8%). Because cul-de-sacs are concentrated in newer, outer urban areas, and tend to be occupied by owner-occupiers on better incomes, while inner urban areas have older through streets, with more renters and lower incomes, social effects can easily be mistaken for spatial effects.

Second, you have to show how each crime is facilitated by spatial design. For example, for burglary, how the burglar accessed the dwelling from public space is critical, not the address. There is no advantage in having an address in a 'safe' cul de sac if the burglar gets in from a footpath behind your back garden. Third, you have to distinguish between different types of crime, and make research crime specific because different crimes relate to space in different ways. Picking pockets thrives in crowded shopping areas, public order offences will concentrate on main streets, since this is where the main pubs and takeaways will be, and so on. Fourth, to establish that some types of location are safer than others you have to establish the rate of crime against the number of targets, not simply show the numbers of crimes.

All of these issues are dealt with in our work. We selected towns with very different social mixes, and sub-areas with both different spatial layouts and different social composition and looked for common patterns in spite of these differences. We focused specifically on domestic burglary in residential areas (and to a lesser extent car crime and vandalism) in order to deal with the type of crime that most affects ordinary people with the exactness that is needed. We mapped the exact part of public space from which the burglar gained access to the crime site, and we established the degree to which different layout factors affect rates of crime in different locations.

Our results are pretty much common sense. They show that the spaces from which you were least likely to be burgled were well integrated linear residential through roads with good numbers of 'intervisible' dwellings facing the road on either side with no gaps. Where through roads did not have these properties they quickly became much more vulnerable. Cul de sacs which did have these properties, that is, were simple, linear and continuously faced with dwellings on both sides, were safer than tortuous complexes of cul de sacs with small segregated groups of dwellings in visually separated spaces. In 'complex cul de sacs', the first line in from the through road - and therefore visible from it - was much safer than the lines after the first corner was turned inside the complex.

Most important of all, our results showed that factors to do with position in the overall layout and those to do with the immediate vulnerability of the dwelling - having a hidden driveway, or an exposed side or back fence - interact, so you could not say that this type of space was safer than that type without taking into account the other factors. All had to be present together to maximise safety. But when all the safety features were in place, then spaces which allowed through movement were safer than those which did not.



New Space Syntax findings on the link between crime and design challenge the widely accepted benefits of 'defensible space'. Space Syntax studies show that streets are made safer by the presence of passers-by, as well as residents. Using Space Syntax, through-movement and 'natural surveillance' potentials can be established early on in new housing projects. Detailed studies of social decline and crime patterns in existing housing layouts are being used to identify problematic areas and understand the factors which increase vulnerability.


Space Syntax studies provide empirical descriptions of current spatial and social structures in housing estates which can be used to measure the likely pedestrian effects of redevelopment proposals and, where relevant, make suggestions for improvements to these which will directly benefit pedestrians of all ages.

Space Syntax analysis of urban areas built before the twentieth century has - despite the heterogeneity of individual case studies - found a number of general shared characteristics. First, the tendency towards larger gridblocks with outward facing dwellings and a clear distinction between public and private space. Second, the creation of simple pedestrian route networks involving relatively few changes of direction between most origins and most destinations. Third, a direct correlation between the accessibility of individual streets (defined mathematically, see Hillier and Hanson, 1984) and the individual levels of pedestrian activity on these streets. Fourth, the inclusion of 'strangers' or non-residents among the residential pedestrian community. Fifth, the sharing of streets by people of all age groups and both sexes.

The design of residential estate layouts in the Twentieth Century has - in many, if not most, cases - turned these rules on their heads. Grid blocks have been downsized, dwelling entrances have been arranged often to suit aspect and solar orientation than any other relationship, and as a result the boundaries between public and private space have been blurred; pedestrian route networks have been complicated - and in some cases become labyrinthine; the link between accessibility and movement rate has weakened; strangers have tended to walk around estates rather than through them since this normally involves fewer changes of direction; and finally, the link between children and adults has weakened since, with the proliferation of spaces brought about by the downscaling and complicating of the urban environment, children have tended to explore isolated areas and the 'natural' link to adult supervision has been lost.

These effects are described by Hillier who details a correlation between the radical spatial design of Twentieth Century housing estates and their largely negative social outcomes (1). The absence of strangers - whose presence is taken for granted by residents of 'traditional' housing areas - and the separation of adults from children - leading to the perception of fear among vulnerable older people who encounter groups of young people as they move through segregated areas of modern housing estates - creates a combined effect which went directly against the regenerative intentions of designers, developers and authorities. In this way, many housing estates which were originally recognised for their design qualities - such as the Maiden Lane Estate in Camden and the Marquess Road Estate in Islington - soon became some of the least desirable places to live (2).

By employing space syntax analysis it has been possible to study individual estates and measure the degree to which the spatial and social characteristics described previously are present. These studies have usually comprised three parts: the computer modelling of spatial networks; the direct observation of pedestrian activity; and the statistical correlation of the twin databases to examine the relationship between spatial design and social outcome.

For example, recent computer modelling studies of the Chalkhill, Maiden Lane and Elm Village Estates in north London show the degree to which each of these housing layouts is segregated from its urban context (Figure 1a/1b, where the colour of each street represents the degree of accessibility and potential pedestrian movement on that street from red = most accessible through orange, yellow and green, to blue = least accessible).

Figures 1c and 1d present the findings of a direct observation study of the Chalkhill Estate where distinctions have been drawn between the use of space by men, women and children. Figure 1c shows how the density of pedestrian activity falls sharply from the streets surrounding the Estate towards the south and east. By contrast, children cluster in spaces towards the heart of the Estate, creating a separation between old and young people. Analysis against the findings of the spatial computer model produces two key results:

* first, that the pattern of adult pedestrian movement is directly related to the degree of spatial accessibility with more accessible streets carrying more pedestrian movement;

* second, that it is the downscaling of urban space which creates the conditions for children to explore the less accessible parts of the estate which adults normally avoid.

Other space syntax studies have shown that the complex configuration of urban space in housing layouts such as the Chalkhill Estate - when combined with the effects of inward-facing dwellings - can create a distinct pattern of fear among pedestrian users, as well as an actual pattern of crime and anti-social behaviour. For example, current space syntax studies in housing estates across the South of England report a direct correlation between the location of criminal behaviour (in this case burglaries) and spatial segregation. Results also indicate that an effective pattern of movement is itself one of the most efficient ways to control crime in housing estates (3).

These findings have led researchers to question the concepts of 'enclosure' and 'defensible space' which have for some time guided the design of housing layouts. The Space Syntax findings support the criticism of other urbanist, such as Jane Jacobs who argued that the city has to abandon design principles which destroy the sense of harmony, safety and control in residential areas.

1 Hillier, Bill "Can Architecture Cause Social Malaise", Chapter 5 of "Space is the Machine"
Cambridge University Press, 1996

2 Hunt Thompson Associates. "Maiden Lane: Feasibility Study" for the London Borough of Camden, 1988

3 Jacobs, Jane. The Death and Life of Great American Cities. Penguin Books, Hammondsworth,1961.


1997 'Housing Design and the Virtual Community' (with M Major, A Penn and Professor B Hillier) Making Cities Livable Conference, Charleston, USA.



CRIME POTENTIALS spatial analysis of the grid, Greets Green, UK


SPACE AND CRIME analysis of crime in an Inner London area, UK


1a HOUSING ESTATE integration analysis, Chalkhill Estate, London, UK


1b HOUSING POCKETS Camden Town, London, UK


1c MOVEMENT RECORDS Chalkhill Estate, London, UK


1D STATIONARY USE Chalkhill Estate, London, UK


Space Syntax 2003