This study has been extracted from a longer paper submitted for the 8th Space Syntax Symposium that took place in Chile during January 2012 (www.sss8.cl). It uses Space Syntax analysis to study a range of issues concerning slums, using examples in the Indian city of Agra and the Saudi Arabian city of Jeddah.
This paper aims to help further develop the spatial understanding of slums and informal settlements. The foundations of this paper lie in the unplanned settlements ofJeddah,Saudi Arabia, where the development of strategies to improve more than 50 unplanned settlements has formed a major component of Space Syntax Limited’s work of the past five years. In late 2010, an opportunity arose to studyAgra, and some of its 350+ slums. This study was carried out remotely to contribute to the work of students in the Architecture of Rapid Growth and Scarce Resource (ARGSR) studio at the London Metropolitan University (LMU), who were based inAgrafor a period of time. This enabled work to be carried out initially to test whether there are consistencies in the spatial properties of each city’s slums.
2.0 Spatial characteristics of slums
In unplanned settlements, including Jeddah’s, a set of spatial characteristics have been identified which appear similar across unplanned settlements; while the edges of the settlements are often “active” (Greene, 2003), their internal spatial structures are almost entirely separate from the wider structure of the city (Karimi et al., 2007). This section starts to examine whether these conditions are also found in Agra’s slum areas.
2.1 Spatial modelling in data sparse situations
To be able to test whether the general characteristics of slums identified by Greene and Karimi et al. are present inAgra, an accurate model is required. In turn this requires accurate physical data of the slum areas which presents a consistent problem.
In Jeddah, the work of Space Syntax Limited has been underpinned by the collection of a vast amount of data by theMunicipalityofJeddah. Work in the city over the past five years also means that qualitative information has been collected through visits to the areas, and discussions with people who grew up in the unplanned settlements before the slum condition developed.
InAgra, the collection of data has been more difficult. The spatial element of the study was carried out through the manual production of an axial map which was analysed as a segment model using Depthmap software. Construction of the map required accurate data forAgraand in particular the slum areas. A local NGO, the Centre for Urban and Regional Excellence (CURE), was able to provide LMU with maps as part of a three year programme they are undertaking to make improvements in slums.
Unfortunately, the accuracy of the map varies; in all cases urban block outlines have been clearly defined and in some cases building footprints are defined. Comparison with Google Earth shows some discrepancies between the two sets of information and many areas are different from recorded – although it must be noted that Google Earth itself does not always use current photographs. Satellite imagery does not get rid of the problem that many slums are formed from complex networks of narrow alleyways, which are very difficult to distinguish from photos. The nature of slums, rapidly growing cities, and the commercial value of up-to-date information, means that these inaccuracies will always exist in traditionally collected survey data (see Figure 1).
Figure 1: Extract of map of Agra and corresponding area in Google Earth – note discrepancies between satellite imagery and manually collected survey data.
The accurate modelling of large slum areas where neither the survey data, nor the resources to collect it exists, presents a major obstacle to the use of Space Syntax in these situations. While the increase in accessibility of GPS technology, and the rise of open source mapping communities such as OpenStreetMap (www.openstreetmap.org), has lead to a major increase in the availability of centreline information for many cities, often the slum areas are still missing.
One solution to the problem of data collection is being piloted by the organisation MapKibera (www.mapkibera.org), in the largest Kenyan slum inNairobi: Kibera. Through a process of community involvement and interaction, the street network has been mapped by the local community using traces collected from GPS devices. Additional information including data such as land uses, services and facilities are added as places of interest, and while the map does not show details of building footprints, it does show the network of publicly accessible spaces.
Figure 2: Google Earth image of Kibera (top) and map of slum area produced through MapKibera project (bottom).
2.2 Spatial characteristics of Agra’s slums
There are over 350 areas which classified as slums inAgra. These areas have been defined on the basis of the UN Habitat definition which identifies any household lacking in one or more of the following characteristics as a slum: access to water or sanitation, security of tenure, durability of housing or sufficient living area (UNCHS, 2003).
Geographically, the slums are distributed evenly across the city, although some patterns emerge: a cluster can be seen towards the centre of the old city adjacent to the Red Fort, and along the Amba Prasad Road, while a series of discrete settlements can be seen distributed at similar intervals along the east-west National Highway 2 (NH2) which defines the peri-urban/rural northern edge of the city (Figure 3).
This pattern of distribution appears consistent with the processes described by Veronique Dupont in the paper “Socio-spatial differentiation and residential segregation inDelhi: a question of scale?” (Dupont, 2002). In this paper, Dupont illustrates how Old Delhi has altered during the last century in terms of social composition and physical degradation: as the better off inhabitants who had traditionally lived close to economic activity moved towards the less congested areas of city, low income groups were left behind to occupy expensive properties. In combination with this change in city centre inhabitant has been the development of industrial areas along major transport axes, which has sparked both formal and informal growth in close proximity to each other on the urban periphery.
Spatially, the slums share similar characteristics of location: while they disappear within what could be described as the background network of residential uses (Hillier, 2009), many of them are located as described by Greene with at least one active edge (Greene, 2003) on some of the highest value global choice routes (10,000m) (Figure 4). This potential exposure to a wider scale of movement provides an economic opportunity at the edge of the area, however this edge condition, and the activity it provides tends to function in isolation from the local structure of the slum
These locations bring with them many benefits which include the proximity to the employment opportunities provided by the industrialisation of these areas. In the case ofAgra, which has a major income source in the form of tourism, some of the most significant monuments (the Red Fort, Taj Mahal, Akbar’s Tomb), lie on the convergences of these routes which offer further potential for economic activity.
The lower income residents benefit from living close to these high choice routes as they focus vehicular movement, including forms of public and shared transportation. Proximity to lower cost transport allows residents to access a wider set of potential employment opportunities across the city, althoughIndia’s complex Caste system may negate many of these as the lower caste groups are discriminated against in such a way that employment opportunities may be affected (Vithayathil et al.2011).
In contrast to the formalised areas of residential use, the slums display a set of spatial characteristics which distinguishes them: While many of the formalised residential areas appear to have a similar lack of accessibility at the global scale, they have been planned on similarly scaled, often discontinuous, rectilinear grids, which do not create the variation in urban morphology that generates a local scale spatial structure. The slum areas, which are self-built, or adapted around ancient monuments and abandoned structures, develop the variations in route and block structure which helps to define local-scale spatial structures (Figure 5).
The definition ofAgra’s slum area boundaries does not appear to have taken spatial structure in to account, and specific boundaries vary widely by area. This means that what may be spatially defined as a single structure could in fact be divided in to two or three slum areas. This discrepancy between spatial and administrative boundaries means that there can be a degree of inaccuracy in using slum boundaries to carry out statistical studies of characteristics.
Figure 4 (left): Choice analysis at radius 10,000m. Agra’s slums often have an active edge but disappear in to the background network of the city. 1 Taj Mahal, 2 Red Fort, 3 Akbar’s Tomb
Figure 5 (right): Choice analysis at radius 400m. While many of the formalised/planned residential areas are still within the background network of the city, the slum areas develop their own structures and centres.
2.3 Comparison to Jeddah’s unplanned settlements
Jeddah’s unplanned settlements – which translate in Arabic as random areas – were defined by theMunicipalityofJeddah, based primarily on identifying changes in the urban morphology between the planned and unplanned areas.
The boundaries of Jeddah’s unplanned settlements are inherently spatial, and it is very rare that a definable local spatial structure is split in to more than one settlement in Jeddah, as happens inAgra.
Jeddah’s 50+ unplanned settlements are also located across the city in a wide range of locations from the historic city centre, to the periphery of the urban agglomeration (Figure 6). The change in these areas that has occurred over the last 30 years due to socio-economic and physical condition shares many similarities with the case ofDelhi.
The unplanned settlement characteristic described by Karimi et al. (2007), whereby the local scale networks that develop within settlements are independent of the hierarchies that develop at the global scale can also be seen inAgra. One of the differences in this characteristic is the degree to which the slums define themselves as separate spatial entities. In Jeddah the difference is startlingly clear at a choice radius of 800m, inAgrathis difference is less clear and occurs at the smaller radius of 400m.
One reason for the difference between the two cities, and the reduced difference between planned and unplanned areas in Agra, could be the variation between typical planned sub-division grids in the two cities: while the unplanned/slum areas have similarly dense spatial networks, in Jeddah the typical shortest edge of a planned urban block is almost uniformly 60m, while in Agra it varies between 25m and 70m but is typically under 50m (Figure 7).
In essence, all residential use inAgrais similarly segregated, although perhaps this should not come as a surprise: many new developments inDelhiseek to segregate themselves through resident screening processes during the sale or rental process, gating communities or employing security guards (Vithayathil et al., 2011). While trends in the housing market must always be viewed carefully as they only reveal patterns about the choices which are available within the market at that time, the following section uses an example from an Agran slum to substantiate the desire for social segregation in residential areas in India.
While the different average block sizes found in each city create small differences between the scales of analyses at which the characteristics are visible, there are many spatial similarities between slums inAgraand Jeddah, most notably the condition of active settlement edges combined with a distinctly different local and global scale spatial structure.
Maurice Mitchell, Robert Barnes and the students of the Architecture of Rapid Growth and Scarce Resource studio at the London Metropolitan University.
Davis, M, 2006, Planet of slums, Verso
Dupont, V, 2002, Socio-spatial differentiation and residential segregation inDelhi: a question of scale? Elsevier
Greene, M, 2003, Housing and community consolidation in informal settlements: A case of movement economy
Hillier, B, 2009, Spatial sustainability in cities: Organic patterns and sustainable forms
Karimi, K, Amir, Abdulgader, Shafei, K, Raford, N, Abdul, E, Zhang, J, Mavridou, M, 2007, Evidence-based spatial intervention for regeneration of informal settlements: the case of Jeddah central unplanned areas
MunicipalityofJeddah, 2009, Jeddah Strategic Plan
UNCHS (Habitat), 2003, The challenge of slums: Global report on human settlements, Earthscan Publication Ltd.London.
Vithayathil, T and Singh, G, 2011, Spaces of discrimination: residential segregation in Indian Cities