Mapping and examining the spatial dimensions of opportunity in 

Ekurhuleni metropolitan area: Exploratory spatial analysis 

 
  

NAME: PRUDENCE MACHEBELE 

STUDENT NO: 756550 

SUPERVISOR: DR GINA WEIR-SMITH  

A research report submitted to the School of Geography, Archaeology and Environmental Studies at the University of 

Witwatersrand, for the fulfilment of Masters of Science in Geographic Information System and Remote Sensing. 

OCTOBER. 2020 

 



i 
 

DECLARATION 

I Prudence Machebele declare that this report is my unaided work. The work was submitted to 

the School of Geography, Archaeology and Environmental studies for the fulfilment of the Master 

of Science in Geographic Information System and Remote Sensing. And has not been submitted 

anywhere for examination. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
X
Date

X
Prudence Machebele

2.October.2020 



ii 
 

DEDICATION 

I dedicate this work to my lovely mom and my two siblings. 

 

  



iii 
 

ACKNOWLEDGMENTS 

First, and most importantly, I would like to thank my Lord and saviour Jesus Christ for the strength 

and grace to be a master of science candidate. Through His eternal love, I managed to face each 

day with a positive mind believing that I am capable of completing this research project. Some 

the days I felt like giving up, but He renewed my strength to press on. At the beginning of this 

research journey, I took for granted that I could work the whole night, until the point where I 

almost broke down. Especially at times when nothing seemed to be working out despite the hard 

work and effort, I was putting. I also appreciate my supervisor, Dr. Gina Weir-smith for the 

support and guidance that she gave me throughout this journey. Last but not least, I would like 

to say ’nakhensa swinene’ to my family and my loved ones particularly my lovely mom for 

supporting and believing in me.  

 

 

  



iv 
 

ABSTRACT  

The South African spatial pattern opportunity continues to be marked by segregation and 

fragmentation even after apartheid. The apartheid regime has created a country that is spatially, 

economically, and socially uneven. Some areas are well-developed while others are marginalised 

and under-privileged. Regardless of the country’s interventions through development 

programmes and the provision of free services, the country’s metropolitan areas remain 

fragmented. Ekurhuleni is one of the metropolitan areas that resemble this fragmentation. 

This research study sought to examine the spatial distribution of opportunity in the Ekurhuleni 

Metropolitan area using exploratory spatial analysis. The data for the analysis were acquired 

from the Census of 2011 conducted by Statistics South Africa, Pick & Pay, Spar, and the 

Department of Health. A total of nine opportunity indicators were used to calculate three 

opportunity dimension scores. The dimension scores were then used to calculate the overall 

opportunity indices for thirty-three main places.  

The findings demonstrated the spatial disparity in the distribution of opportunity across the 

Ekurhuleni Metropolitan area. Most of the main places that are known to be towns have good 

opportunities, while poor opportunities are predominantly in townships. Some of the townships 

were also classified as high opportunity areas, however, it does not look aesthetically pleasing. 

There are areas with a high number of people and good opportunity, while others had a high 

number of people and poor opportunity. This brought an understanding that the total number 

of people does not have a clear relationship with the opportunity score. The study concludes that 

there is an uneven distribution of opportunity in the metropolitan area and good opportunities 

are clustered in developed towns, while townships tend to have poor opportunities.  

  



v 
 

TABLE OF CONTENTS 

 
DECLARATION .................................................................................................................................... i 

DEDICATION ...................................................................................................................................... ii 

ACKNOWLEDGMENTS ....................................................................................................................... iii 

ABSTRACT ......................................................................................................................................... iv 

TABLE OF CONTENTS .......................................................................................................................... v 

LIST OF FIGURES .............................................................................................................................. viii 

LIST OF TABLES ................................................................................................................................. ix 

ACRONYMS ........................................................................................................................................ x 

Chapter 1: Introduction and Background ............................................................................................1 

1.1 General Introduction .................................................................................................................... 1 

1.2 Background of the study .................................................................................................................... 2 

1.3 Problem statement ............................................................................................................................ 3 

1.4 Aim ...................................................................................................................................................... 4 

1.5 Objectives ........................................................................................................................................... 4 

1.6 Contribution to knowledge ................................................................................................................ 4 

1.7 Overview of the Report ...................................................................................................................... 5 

CHAPTER 2: DRAWING FROM THE LITERATURE ...................................................................................6 

2.1 Introduction ......................................................................................................................................... 6 

2.2 Dynamics of the geography of opportunity .......................................................................................... 6 

2.2.1 The hypothesis of spatial mismatch theory ................................................................................... 7 

2.2.2 Neighbourhood segregation .......................................................................................................... 9 

2.3 Livelihoods perspective ...................................................................................................................... 11 

2.4 Application of exploratory spatial data analysis .................................................................................. 13 

2.4.1 Rationale ..................................................................................................................................... 13 

2.4.2 Methodological approaches to ESDA ........................................................................................... 14 

2.5 Conclusion ......................................................................................................................................... 15 

CHAPTER 3: RESEARCH METHODOLOGY .............................................................................................. 17 

3.1 Introduction ....................................................................................................................................... 17 

3.2 Background of the Study site ........................................................................................................... 17 

3.3 Data Acquisition ............................................................................................................................... 18 



vi 
 

3.4 Data analysis ..................................................................................................................................... 20 

3.4.1 Data pre-processing .................................................................................................................. 20 

3.4.2 Descriptive statistics on the raw data ...................................................................................... 21 

3.4.3 Creation of the Opportunity Index ........................................................................................... 21 

3.4.4 Choropleth Mapping ................................................................................................................. 22 

3.4.5 The Neighbours and spatial weights matrix ............................................................................. 22 

3.4.6 Spatial autocorrelation ............................................................................................................. 22 

3.4.7 Hot spot and cold spot analysis ................................................................................................ 23 

3.5. Limitations ....................................................................................................................................... 23 

CHAPTER 4: RESULTS ........................................................................................................................ 25 

4.1 Introduction ...................................................................................................................................... 25 

4.2 Descriptive statistics ........................................................................................................................ 25 

4.3 Opportunity Index ............................................................................................................................ 27 

4.4 Mapping the distribution of opportunity ........................................................................................ 30 

4.4.1 Opportunity Index choropleth map .......................................................................................... 30 

4.4.2 Jobs and local economy choropleth ......................................................................................... 37 

4.4.3 Education ................................................................................................................................... 38 

4.4.4 Community health and civic life ............................................................................................... 40 

4.6 Examining the spatial pattern of opportunity ................................................................................. 42 

4.6.1 Local Moran ............................................................................................................................... 42 

4.6.2 Hot spot and cold spot clusters ................................................................................................ 43 

4.7 Conclusion ........................................................................................................................................ 44 

CHAPTER 5: DISCUSSION OF FINDINGS ................................................................................................ 45 

5.1 Introduction ....................................................................................................................................... 45 

5.2 Reflection on the significance of the main findings ............................................................................ 45 

5.2.1 Descriptive statistics .................................................................................................................... 45 

5.2.2 Opportunity ................................................................................................................................ 47 

5.2.3 Local Moran ................................................................................................................................ 48 

5.3 Reflections on the findings in relation to the literature ...................................................................... 48 

5.3.1 The hypothesis of spatial mismatch theory ................................................................................. 48 

5.3.2 Neighbourhood segregation ........................................................................................................ 49 

5.3.3 Livelihoods perspective ............................................................................................................... 50 

5.4 Conclusion ......................................................................................................................................... 50 



vii 
 

CHAPTER 6: Conclusion ...................................................................................................................... 51 

6.1 Summary and research overview ....................................................................................................... 51 

6.2 Recommendations ............................................................................................................................. 52 

6.3 Personal reflections............................................................................................................................ 53 

6.4 Conclusion ......................................................................................................................................... 53 

References ....................................................................................................................................... 54 

Appendix A: Total population per population groups ........................................................................ 60 

Appendix B: Metadata for point attribute data ................................................................................. 61 

Appendix C: Population groups per main places ................................................................................ 62 

 

 

 

  



viii 
 

LIST OF FIGURES 

Figure 1:Map showing the location of Ekurhuleni metropolitan area(Author,2020) .................. 17 

Figure 2:Opportunity index choropleth map(Author,2020) ......................................................... 31 

Figure 3:Opportunity index choropleth with total population distribution(Author,2020) .......... 32 

Figure 4:Image showing some parts of Boksburg town( Van Zyl,2013) ....................................... 33 

Figure 5:Image showing the recreational area in Boksburg( Wang,2018) ................................... 34 

Figure 6:Image showing OR Tambo International airport( TimesLIVE,2017) ............................... 34 

Figure 7:The physical appearance of some parts of Katlehong  township (Janine, 2018 ) .......... 35 

Figure 8:Location of school and shacks along Peterson road in Dukathole (Street view,2017) .. 36 

Figure 9:Cluster of shacks and informal trading in Thokoza township (Khumalo, 2014) ............. 36 

Figure 10:choropleth map for jobs and local economy dimension(Author,2020) ....................... 37 

Figure 11:Choropleth map for Education(Author,2020) .............................................................. 39 

Figure 12:School in Boksburg (Grove,2010) ................................................................................. 39 

Figure 13:Choropleth map for community health and civic life dimension(Author,2020) .......... 40 

Figure 14:Grocery stores in Duduza township(Baloyi,2013) ........................................................ 41 

Figure 15:Street vendors in Thokoza township(Andrews,2009) .................................................. 41 

Figure 16:Local Moran statistic map(Author,2020) ...................................................................... 42 

Figure 17:Hot and cold spot map(Author,2020) ........................................................................... 43 

 

 

  

file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612831
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612832
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612833
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612834
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612835
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612836
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612837
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612838
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612839
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612840
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612841
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612842
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612843
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612844
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612845
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612846
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47612847


ix 
 

LIST OF TABLES 

Table 1:datasets and their sources ............................................................................................... 20 

Table 2:Descriptive Statistics for Jobs and local economy indicators .......................................... 26 

Table 3:Descriptive Statistics for Community health and Civic life indicators ............................. 26 

Table 4:Descriptive Statistics for Education indicators ................................................................ 27 

Table 5:Opportunity Index ............................................................................................................ 28 

Table 6:Descriptive statistics of the opportunity index ................................................................ 29 

Table 7:Rental expenditure statistics ........................................................................................... 30 

  

file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47613768
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47613769
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47613770
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47613771
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47613772
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47613773
file:///F:/MASTERS_NEW/RESEARCH/Report/Report.docx%23_Toc47613774


x 
 

ACRONYMS 
 

CBD             Central Business District 

CoE              City of Ekurhuleni 

EMA            Ekurhuleni Metropolitan Area 

ESDA           Exploratory Spatial Data Analysis 

GIS               Geographic Information System 

GCRO          Gauteng City Region Organisation   

StatsSA         Statistics South Africa     



1 
 

Chapter 1: Introduction and Background  

1.1 General Introduction  
 

The disconnection between communities in the metropolitan area has been dealt with by the 

studies of the geography of opportunity (Powell et al., 2007). The spatial dimension of 

opportunity also known as the ‘geography of opportunity’ refers to the spatial dimension of 

opportunity at a metropolitan level. It explains that the social and spatial outcomes of a 

neighbourhood are shaped by the interactions between the institutional system and people’s 

choices (Wilson and Greenlee, 2016). Opportunity is not a synonym of jobs; however, it includes 

dimensions such as jobs and local economy, education, community health and civic life that 

enables monitoring of the neighbourhoods (Wilson and Greenlee, 2016). The local economy 

includes indicators such as unemployment rate, poverty rate, level of income, and access to the 

internet while education includes the level of educational attainment and dropout rates. Health 

and civic life include access to health services and people’s involvement in social activities (Wilson 

and Greenlee, 2016).  

The spatial distribution of opportunity in different countries like South Africa, United States 

regions has proven to be racialized and has created the notion of two cities in one. The two cities 

show a distinction between areas where whites and blacks stay (Reece et al., 2009; Powell et al., 

2007; Howell-Moroney; 2005). Most of the affluent places were predominantly white, while 

black people tend to stay in vulnerable parts of these countries (Powell et al., 2007). The 

privileged communities have advanced job opportunities, schools with well- developed 

education systems and grocery stores that sell healthy goods. Poor communities are 

characterised by factors that include high crime levels, marginalisation, poor resources, 

unemployment and the concentration of poverty. 

According to Mahajan (2014), South Africa is currently known for uneven development and 

spatial gaps, which have also influenced the spatial distribution of opportunities in the entire 

country. Some of the country’s regions are developed while others are lagging (Peberdy et al., 

2017). Development programmes have been adopted by both local and provincial governments 

to combat the spatial imbalances (Venter, 2016) however, the country remains spatially uneven.  



2 
 

The townships within the country accommodate most of the country’s population, and these 

townships are at the peripheral locations where there are limited and poor initiatives (Peberdy 

et al., 2017). People in these areas mainly depend on informal employment which negatively 

affects the growth of their local economies (Peberdy et al., 2017). This research set out to 

examine the geography of opportunity in the main places of the Ekurhuleni Metropolitan area 

and reveal how it impacts people's livelihoods.   

 

The study used opportunity indices to conduct the analysis. The indices were created based on 

census data, Department of Health and retail stores data. According to Opportunity Nation 

(2017), the Opportunity Index is an annual composite measure of economic, educational and civic 

factors at the state and county levels. The Opportunity index measures the access that people 

have to opportunity in their neighbourhoods across the metropolitan area. It is measured using 

three dimensions namely, Jobs and local economy, Education, Community health, and civic life 

as indicated by the study of Wilson and Greenlee (2016). 

1.2 Background of the study 
 

The geography of opportunity embraces a wide range of themes that include, but are not limited 

to spatial mismatch, uneven development, neighbourhood segregation and spatial inequalities. 

These themes explain a particular dimension of opportunity. Due to uneven urban structures, 

most of the themes consider the uneven distribution of opportunity in metropolitan areas of 

both international and local contexts. 

 

This research report presents the findings of the spatial distribution of opportunities in the 

Ekurhuleni metropolitan area. The emphasis is on determining the clusters of both low and high 

opportunity and to understand the distribution of spatial opportunity of the metropolitan area. 

The author of this research study contends that there is an uneven distribution of opportunity in 

the metropolitan area. Well-advanced opportunities are located in white-dominated areas, while 

black people are predominant in poor opportunity areas. Therefore, as a result, there is a 

noticeable distinction between the livelihood outcomes of people in both geographic locations. 



3 
 

To improve the uneven distribution of opportunity, interventions such as place-based 

developments should be adopted to target the development of each neighbourhood.  

 

1.3 Problem statement  
 

The spatial arrangement of the opportunities is significantly interlinked to the economic structure 

and the livelihoods of people within the area. The arrangement also indicates the spatial patterns 

of how the opportunities are located in the neighbourhoods (Picard and Zenou, 2018). Ekurhuleni 

metropolitan area is characterized by spatial disparities and unevenly developed places (Todes 

et al., 2010). These places display spatial gaps through the location of advanced amenities within 

the economic nodes while surrounded by overpopulated townships at the outskirts (Todes et al., 

2010). 

Ekurhuleni Metro is one of the three metropolitan areas located in Gauteng. The province is one 

of the richest and densely populated in the province with about 3.2 million population and 896 117 

households (Marutlulle, 2017). The Ekurhuleni Metropolitan area has a very diverse economy and 

accounts for almost a quarter of Gauteng’s economy (Marutlulle, 2017). The metropolitan area 

is traditionally known as a manufacturing heartland and accommodates most of the country’s 

factories of production (Cross et al,2005; Todes et al., 2010; Marutlulle, 2017). The industrial 

sectors provide jobs for both skilled and semi-skilled workers (Cross et al,2005). The municipality 

is also home to the OR Tambo International Airport, which is a transport hub that services both 

international and local travellers (City of Ekurhuleni, 2019; Marutlulle, 2017). Other 

developments (such as retail development, public sector investment) are also emerging and 

growing the economy and attracting people into the metropolitan area (Rogerson, 2018; Todes 

et al., 2010). 

 

The metro has the fastest growing population in the country with most of the households 

spending their income in rental housing (Stats, 2011). The influx of people is causing a high 

constant demand for housing because most of the households do not have formal shelters 

Marutlulle, 2017). According to Marutlulle (2017), Cross et al (2005) and Mushongera (2018) high 



4 
 

influx of people into the metropolitan area has resulted in a concentration of informal 

settlements and induced poverty – making this metropolitan area suitable for the study. 

Theoretical work has been done to analyse the geography of opportunity at the international, no 

official research has been done in South Africa, nor Ekurhuleni, about geography of opportunity. 

Some studies have been done dimensions other than all aspects that determine the geography 

of opportunity. For instance, there is much work done on issues such as poverty, informality, and 

unemployment (Peberdy et al., 2017; Marutlulle, 2017; Cross et al,2005 and Mushongera, 

2018).Some tackled the spatial mismatch and neighbourhood segregation (Naudé, 2008; Turok 

et al., 2017). However, there has been a minimal focus by scholars on the spatial distribution of 

opportunity in South Africa. This research adds to the existing body of knowledge by focusing on 

the spatial dimension of opportunity in the metropolitan area. It focuses on the collective 

dimensions that explain the geography of opportunity in the metro. 

1.4 Aim  

The main purpose of the study was to assess the spatial distribution of opportunities by 

identifying clusters of high and low opportunity in the Ekurhuleni metropolitan area.  

 

1.5 Objectives  
The research focussed on the following objectives:  

• To create an opportunity index using the three opportunity dimensions which are, jobs 

and local economy, education, community health, and civic life;  

• To map the results from the opportunity index; 

• To identify the spatial patterns of opportunity using exploratory spatial analysis tools. 

1.6 Contribution to knowledge  

The theory of the geography of opportunity has been of interest to international authors. have 

tackled various factors of the geography of opportunity using exploratory analyses. However, 

none of the studies were done to understand the spatial dimension of opportunity and more so 

in a local context like the Ekurhuleni metropolitan area. Therefore, this research report will be 



5 
 

the first to present knowledge about the clustering of opportunities within the metropolitan. This 

study will inform officials about the sense of where to target for development purposes. It will 

also give direction to future researchers to expand beyond understanding the spatial dimension 

of opportunities in metropolitan areas to a broader context. 

 

1.7 Overview of the Report 

The first chapter of the research report presents the background and outlines the scene of the 

study. The background leads to the introductory chapter that summarises the purpose of the 

study. The second chapter presents the literature that supports the research topic. This literature 

review touches on some of the conflicts, consensus, including the gaps that give a clear 

understanding of the study focus. Chapter three of the research report presents the contextual 

setting of the study area and why it is suitable for this research. The chapter outlines the 

methodological approaches of how the data has been acquired and analysed. The data 

acquisition approach includes requesting data from various institutions and departments, while 

data analyses include approaches such as creating the opportunity index and examining the 

spatial patterns of opportunity. Chapter four presents the results found after performing the 

analysis. This findings chapter, together with the literature review builds the arguments 

presented in chapter five. The arguments in chapter five also enable the achievement of the 

research's aim and objectives. The last chapter is the concluding remarks and recommendations. 

This chapter reflects on every chapter of the research to draw an informed conclusion. 

 

 

 

 

 



6 
 

CHAPTER 2: DRAWING FROM THE LITERATURE 

2.1 Introduction 
There has been empirical work done to explain the spatial dimension of opportunity in 

underdeveloped metropolitan regions (Lens, 2017; Powell et al., 2007; Hellerstein et al., 2008; 

Picard and Zenou, 2018; Howell-Moroney, 2005). Although the literature on this topic includes 

different theories, this report will only be based on the major themes which emerged repeatedly 

in the literature. The three major themes include the dynamics of the geography of opportunity, 

livelihoods perspective, and the exploratory spatial data analysis. The geography of opportunity 

literature will draw on mechanisms such as spatial mismatch and neighbourhood segregation. 

Exploratory spatial data analysis will focus on the rationale and methodological aspects. Most of 

the literature presents the international contexts of these themes and few are based on South 

African contexts. The specific literature that relates to the study include; Wilson and Greenlee 

(2016), Howell-Moroney (2005), Naudé (2008), and Powell et al (2007). This literature presents 

an understanding of the geography of opportunity in both international and South African 

contexts. 

 

2.2 Dynamics of the geography of opportunity  
 

The geography of opportunity is a phrase that refers to the “spatial dimensions of opportunity at 

the metropolitan level” (Wilson and Greenlee, 2016; 626). It explains that the structural aspects 

of the neighbourhood and people’s preferences influence the social, economic, and spatial 

prospects (Brain and Prieto, 2018; Galster and Killen, 1995; Wilson and Greenlee, 2016). It has 

been referred to as communities of opportunity by Powell et al (2007), where it implies that the 

area where people are located within the metropolitan area determines their access to 

opportunity. In support of Powell’s argument, Picard and Zenou (2018) argue that the 

metropolitan region that is developed should improve access to opportunity and livelihood of all 

the residents, while the underdeveloped one is characterised by unequal distribution of 

opportunity.  



7 
 

The notion of developed and underdeveloped areas in one metropolitan area is predominant in 

most countries. In these instances, advanced opportunity is clustered in the white-dominated 

areas while, poor living environments characterises the areas dominated by black people 

(Howell-Moroney, 2005; Powell et al, 2007; Hellerstein et al, 2008; Picard and Zenou, 2018). This 

is the case for most South African metropolitan areas not only Ekurhuleni where white people 

are predominant in well-developed areas while black people stay in marginalised townships 

(Todes et al., 2010; Naudé, 2008). 

 

 

2.2.1 The hypothesis of spatial mismatch theory 
 

The uneven spatial distribution of opportunity has been discussed in a variety of literature with 

theories such as spatial mismatch and residential segregation (Naudé, 2008; Howell-Moroney, 

2005; Gobillon et al., 2007). The spatial mismatch refers to the physical separation between job 

opportunities and residential areas. It can be explained by three spatial restrictions namely: 

space, race, and job information (Howell-Moroney, 2005; Naudé, 2008). Space restriction theory 

explains that long commuting times restrict access to a job opportunity. Turok et al., 2017 and 

Gobillon et al (2007) argue that outcomes for residents that are spatially disconnected from the 

job opportunities are associated with the poor labour market.  

 

The housing market is another factor that causes space restriction (Hellerstein et al., 2008; Picard 

and Zenou, 2018; Squires and Kubrin, 2005). The job opportunities in most metropolitan areas 

have been decentralised to suburban areas where most of the white people have also moved to, 

hence the increase in housing prices (Howell-Moroney, 2005). Housing markets are determined 

by land prices, and it has been observed that in areas with high land prices there are good 

opportunities, while low land price areas are characterised by poor opportunity (Brain and Prieto, 

2018). They also argue that low-income housing is in underprivileged areas while expensive 

housing markets are in economically advanced areas. In the South African context (particularly 

in Gauteng), the apartheid legacies have resulted in poor housing and informal settlements 



8 
 

located at the peripheral edges where black people stay (Peberdy et al., 2017). Although the post-

apartheid government has adopted policies to redress these challenges; however, the housing 

services that are provided indicate marginality and peripherality (Peberdy et al., 2017). For 

instance, the RDP (breaking new ground) housing developments are targeted at the peripheral 

areas away from the major economic developments. 

The race restriction explains that jobs provided at proximity to some of the residential areas cater 

for white people rather than black people (Howell-Moroney, 2005; Gobillon et al., 2007. In 

support of their arguments, Hellerstein et al., (2008) contend that spatial mismatch is not always 

the space restriction (absence of jobs in areas where black people are located), but the lack of 

jobs in which black people could be hired (labour market). One could agree with the authors’ 

arguments about space and race restriction particularly regarding South Africa because the job 

opportunities are available in proximity to black townships, however, black people do not have 

sufficient skills to participate in those job opportunities. Therefore, they are forced to commute 

long distances for jobs requiring low skills.  

According to Naudé (2008), the friction of distance is another cause of racial mismatch and is 

experienced differently by both black and white groups. The friction of distance, according to 

Naudé (2008) refers to the challenges associated with the distance between job opportunities 

and residential areas. This implies that there are different experiences between blacks and whites 

concerning commuting costs or access to transport. These factors have noticeable implications 

on employment status and the level of income for black people (Naudé, 2008). One could agree 

with Naudé; for example, some black people spend considerably more time in public transport 

to access their workplaces. Moreover, they also spend most of their limited salaries on public 

transport. Therefore, one could also argue that spatial mismatch resembles the racial mismatch 

in the context of South Africa. 

 

Job information restriction theory explains that black people have limited information about the 

informal way of job applications (Howell-Moroney, 2005; Gobillon et al., 2007). According to 

Gobillon et al. (2007), the diversity of the network that a person has can also determine the job 



9 
 

information they will have. In a sense that a person that moves beyond their immediate circle, 

share strong networks, and non-redundant job information. This is also applicable in the South 

African context because Gobillon et al. (2007) indicate that unemployment is induced by 

inappropriate means of searching for a job opportunity. Moreover, most black people live in 

segregated locations where they tend to acquire information only from the people that they 

socialise with daily (Peberdy et al., 2017). 

The outcome of the spatial mismatch differs from geographical contexts. In the case of the US 

counties, the inner cities are characterised by unemployment and poor opportunity and are 

mostly occupied by black people (Glaeser et al., 2004; Hellerstein et al., 2008; Gobillon et al., 

2007). In the South African context, black people are mainly located in the peripheral townships 

away from the city centres. However, this trend is changing because inner cities are becoming 

more occupied by black people (Peberdy et al., 2017). The spatial mismatch arguments relate to 

the case of the Ekurhuleni Metropolitan area, Todes et al (2010), outline that as a result of 

apartheid legacy, black people remain located in the peripheral townships that are characterized 

by high unemployment rates. Areas where they have to commute long hours to access 

opportunities within the metropolitan area (Naudé, 2008). Thus, it indicates that there is an 

uneven distribution of opportunity in the metropolitan area to a certain extent that will be 

substantiated by this study. 

 

2.2.2 Neighbourhood segregation  
 

The neighbourhood segregation includes the separation of individuals from livelihood 

opportunities based on race, class, and income (Lens, 2017; Venter et al., 2007). It is also referred 

to as the segmentation that creates neighbourhood disparities and has an impact on people’s 

welfare, such as level of education and poverty (Howell-Moroney, 2005). This happens because 

the metropolitan's economy is concentrated in specific geographic areas which results in a high 

level of unemployment in marginalised communities (Turok et al., 2017). Marginalised 

neighbourhoods also have a poor economic structure and labour market which restricts access 

to employment opportunities (Powell et al., 2007; Howell-Moroney, 2005). In support of the 



10 
 

argument, Márquez et al, (2019) indicate that the geographic location of the neighbourhood has 

an impact on its economic structure. In these communities, grocery stores sell poor quality goods 

at a higher price (Powell et al., 2007). Residents in the segregated neighbourhoods stay without 

work for a longer period because of the lack of job opportunities in their communities and 

inadequate information about job opportunities. One could argue with the statement because 

most residents of similar employment status are likely to stay together, and they only interact 

with unemployed individuals within their segregated neighbourhoods (Picard and Zenou, 2018). 

However, if they were integrated, it would enable the sharing of information through different 

platforms. One could contend that segregation makes unemployment pervasive in some of the 

townships because the interaction is limited to the neighbourhood. 

 

According to Lens (2017), neighbourhoods are not just segregated by race, class, and income, but 

they also have an impact on an individuals’ livelihood outcome. Howell-Moroney (2005) and 

Powell et al (2007) also gave an instance where educational outcomes of students in segregated 

communities are poor (with more drop-outs) compared to those that are in integrated 

neighbourhoods. Hellerstein et al (2008) did their study in cities like Chicago in the United States, 

which outlined the distinctive difference between the educational level of the blacks and the 

whites. It was found in their study that the segregation the black communities is the main 

influence of the differences in education levels. Their argument is also applicable to the context 

of the South African neighbourhoods where Naudé (2008), Howell-Moroney (2005) and Venter 

et al (2007) indicate that the level of educational attainment of people in poor black 

neighbourhoods, is lower compared to those in well-advanced communities.  

 

These poor communities are also associated with high dropout rates as compared to white 

communities. The condition of segregated neighbourhoods also has a major influence on the 

concentration of poverty, and the quality of institutions such as schools and hospitals 

(neighbourhood effects) (Turok et al., 2017; Powell et al., 2007). In these neighbourhoods, the 

institutions are under-resourced and inadequate to meet people’s needs (Powell et al., 2007; 



11 
 

Howell-Moroney, 2005). Neighbourhoods that are well-developed enhance the livelihoods of the 

residents while poor neighbourhoods do not (Turok et al., 2017). 

 

The mismatch of space and residential segregation are closely related (Hellerstein et al., 2008). 

This is because spatial mismatch could be a result of residential segregation. One could agree 

with the argument because the more people are segregated spatially, the more they are 

separated from opportunities such as employment. In the case of the Ekurhuleni Metropolitan, 

Todes et al (2010) indicate how peripheral townships are characterised by high unemployment 

because they are disconnected from the economic centres.  

 

 

2.3 Livelihoods perspective 
 

The geography of opportunity has also been used to understand the difference between the 

livelihoods of residents in poor and well-advanced neighbourhoods (Howell-Moroney, 2005; 

Powell et al., 2007). The concept of livelihood explains the way people live in their 

neighbourhoods which is determined by the assets that enable one’s everyday living (Rakodi, 

2002). Before modernisation, the literature about livelihood perspective was centered on rural 

development in which factors such as, wage employment, farm labour, and agriculture were used 

to understand the collective livelihoods of the entire community (Scoones, 2009). However, the 

rise of the modernisation era has shifted the focus to the understanding of livelihoods in the 

urban context (Scoones, 2009; Rakodi, 2002).  

For this study, understanding urban livelihoods' perspective would be relevant, because it has 

direct relations with the spatial location of opportunity. Livelihood refers to all the assets 

essential for the means of survival (Rakodi, 2002), These are assets grouped into five categories 

based on their functions for an individual’s living standard. The five categories include physical 

capital, social and political capital, human capital, natural capital, and financial capital. According 

to Rakodi (2002), political and social capital are social resources that promote social inclusion and 

a sense of belonging. These resources include community meetings, membership groups, and 



12 
 

other institutions. Social and political resources create a platform for residents to share 

information and raise issues that affect their livelihoods. Human capital refers to the labour 

resources that enable one to be competent for job opportunities. Labour resources include 

educational and non-educational skills. These livelihood assets allow all individuals to generate 

income based on the skills they have. For instance, one could generate income from their 

businesses while others could rely on a professional job to generate income. Natural capital is an 

all-natural resource that people can utilise to produce goods and services. Some businesses use 

primary goods such as trees to produces secondary goods like furniture. The financial capital 

refers to savings or pensions. These resources enable the exchange for goods and services. Lastly, 

physical resources include infrastructures including houses and transport. The basic 

infrastructures ensure safety and security (Rakodi, 2002). 

Studies have been done to understand the impact of livelihood resources. Chen et al (2016) gave 

an instance of the financial capital (particularly addressing the informal employment). In their 

study, they argue that in most developing countries, people generate their income through 

informal employment such as street vendors and road-side shops, etc. As a result, people’s 

livelihoods are affected because informal employment is exposed to various threats. For 

instance, some government laws and regulations limit the growth of informal employment by 

regulating where these informal sectors should locate. Hovsha and Meyer (2015) also give a 

similar scenario for the South African context. In their research, they argue that due to a lack of 

educational qualifications and skills, informal employment has a huge impact on the livelihoods 

of South Africans. Although there are municipal by-laws that regulate informal employment, in 

practice, it is not prioritised as compared to formal employment which makes informal 

employment less profitable to benefit business owners. 

In another study Skinner (2000) explains the impact of physical assets on people’s livelihoods. He 

gives instances of energy infrastructures. In the study, he argues that energy is a physical capital 

that is used for purposes like lighting and cooking. Both low (mainly blacks) and middle/high 

(mainly white) income groups have different sources of energy. Low-income groups depend on 

fuels such as paraffin and coal as the main source of energy which exposes their lives to fire 

hazards (Skinner, 2000). According to GCRO (2017/2018), several households in Ekurhuleni 



13 
 

metropolitan townships still depend on energy sources such as woods, candles, and paraffin. This 

indicates that poverty is a social issue within the metropolitan area.   

Transportation is another physical asset. Sohail (2000) indicate that low-income people mainly 

use public transport for their mobility. Most of these public transports are inadequate and not 

reliable. However, due to high commuting costs, they are forced to settle for these types of 

transport. For instance, some of the low-income groups in South Africa use Metrorail trains and 

other public transport to commute for about 20 km, this forces them to spend most of their time 

away from their families. With regard to this argument, it is clear that public transport has a 

negative effect on the livelihood of people. The GCRO (2017/2018) also indicates that numerous 

households in the Ekurhuleni metropolitan area use public transport to commute to their 

workplaces. Most of these households travel long distances in their everyday lives (Naudé, 2008). 

Thus also indicating there is a relation between the location of opportunity and people’s 

livelihoods. 

 

2.4 Application of exploratory spatial data analysis  

2.4.1 Rationale  
 

The exploratory spatial data analysis (ESDA) is a research method that is considered as flexible 

and not limited to a specific formal structure (Davies, 2011). It uses tools to visualize, display, and 

to determine patterns of clusters and spatial regimes (Dall’erba, 2003; Celebioglu and Dall’erba, 

2009). These tools are mainly applicable to small regions and take into consideration the 

existence of both spatial heterogeneity and spatial autocorrelation (Rusche, 2008). Spatial 

autocorrelation is taken from Tobler’s first law which explains the relations of the things that are 

close to each other than those that are at a distance. For instance, the regions that are close to 

each other might have the clustering of activities with similar functions. The spatial heterogeneity 

is the opposite of the spatial autocorrelation which explains that effects in the same area could 

have a different extent of impact (Rusche, 2008). The ESDA can analyse the above-mentioned 

two spatial effects, and incorporate them into different contexts to achieve the following 



14 
 

objectives: describing spatial distributions, assessing the spatial patterns, predicting the spatial 

regimes, and identifying possible observations such as outliers (Rusche, 2008). 

Most international studies have dealt with the use of exploratory spatial analysis for different 

purposes (Wang et al., 2016; Anselin et al., 2006; Duque et al., 2013). For example, in countries 

like China, the exploratory spatial analysis has been used as a tool to describe spatial inequalities 

and redistribution of services (Duque et al., 2013; Wang et al., 2016). The analysis process 

involved analysing factors such as unemployment rate and regional economic growth (Duque et 

al., 2013; Wang et al., 2016). In the state of Virginia, the exploratory spatial analysis was used to 

analyse social indicators such as the spatial clusters of high child risk scale (Anselin et al., 2006). 

However, in this research study, the exploratory spatial analysis will be used to identify clusters 

of high and low opportunities in the Ekurhuleni Metropolitan area.  

2.4.2 Methodological approaches to ESDA  
 

Descriptive statistics  
 
Descriptive statistics is the principal component of the exploratory study that provides 

summaries of the raw data, and also shows relationships between the variables (Kaushik and 

Mathur, 2014; Sharma, 2019). This analytical method can be presented numerically or graphically 

(Kaushik and Mathur, 2014). The descriptive statistics can also be utilised to detect outliers and 

normality of the distribution.  

Neighbours and spatial weights  

Spatial neighbours and weights determine the spatial relations between the spatial units (Chen, 

2012; Cheng et al., 2014; Shen, 1994). Spatial analysis is based on the assumption that close 

phenomena are more similar than those that are farther apart (Foster and Evans, 2008). 

Therefore, it is crucial to define neighbours and spatial weights before checking the presence of 

spatial patterns. There are different types of spatial weights, but for this study, only Distance-

based weight will be used. Distance-based weight is the spatial weight that could be measured 

using the distance between two spatial neighbours. It is also called Euclidean distance. It can also 



15 
 

be measured using the distance between the centroids of polygons which is also called Great 

circle distance (Waller and Gotway, 2004; Bivand et al., 2008). 

 

 Choropleth mapping 

The Choropleth map is one of the ESDA tools which are used to display the spatial attributes 

(Leonowicz, 2006). It enables one to compare the differences between the neighbourhoods and 

analysing spatial distributions (Celebioglu and Dall’erba, 2009). 

Spatial autocorrelation 

The spatial autocorrelation measures the degree to which observations at nearby spatial 

locations are similar or dissimilar (Chi and Zhu, 2008). It includes techniques such as global and 

local spatial autocorrelation. The global spatial autocorrelation (known as Moran’s statistic) 

provides the linear relationship between the spatial units. The local l autocorrelation (local 

Moran) can be used to identify the spatial observations that might indicate homogeneity 

(Sridharan et al., 2007; Wang et al., 2016; Arun, 2013; Foster and Evans, 2008). Spatial 

autocorrelation analysis is a crucial step to consider before any analysis of the spatial 

phenomena. The ESDA can also be used to measure the spillover effects across the area that one 

is analysing. Therefore, for this research, only the local spatial autocorrelation (Local Moran) will 

be used because the study site is on a local scale. 

2.5 Conclusion 

This chapter presented the literature using three themes, namely the dynamics of the geography 

of opportunity, livelihood, and application of the exploratory spatial data analysis. The chapter 

has also presented gaps, agreements, and conflicts in the literature. The dynamics of the 

geography of opportunity theme was based on theories such as spatial mismatch and 

neighbourhood segregation. Spatial mismatch theory indicated how the distance between job 

opportunities and place of residence differs between poor and well-advanced neighbourhoods 

and how it affects people’s livelihoods. The neighbourhood segregation outlined how segregated 

communities tend to have a poor opportunity outcome as compared to integrated communities. 



16 
 

The livelihood theme presented all the factors that determine people’s livelihoods, and how they 

are affected by the spatial location of opportunity. Lastly, the exploratory data analysis theme 

outlined the rationale and the methodological approaches of carrying out an exploratory spatial 

analysis. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



17 
 

CHAPTER 3: RESEARCH METHODOLOGY 

3.1 Introduction 
The methodology section outlines the procedure of how the data was acquired, pre-processed, 

and analysed. This research study is exploratory, which is considered as an approach of 

investigating, identifying and manipulating spatial data (Sridharan et al., 2007; Lincaru et al., 

2016; Jung and Vijverberg, 2019), it is flexible and not limited to a specific formal procedure 

(Davies, 2011). 

3.2 Background of the Study site 
 The study is based in the Ekurhuleni Metropolitan area, one of the three metropolitan 

municipalities in Gauteng. It was officially constituted to be a standalone metropolitan 

municipality in 2001 after being part of the Johannesburg metropolitan before (Naudé, 2008). 

The metropolitan covers an area of about 1901.47 km2 (Todes et al., 2010). It is comprised of nine 

towns. Spatially, the area is highly fragmented, consisting of developed suburbs and lagging  

 

 

 

 

 

 

 

 

 

 

 

EMA 

Figure 1:Map showing the location of Ekurhuleni Metropolitan area (Author, 2020) 

 



18 
 

townships (Todes et al., 2010). To give a general overview of the metropolitan's social-economic 

dynamics, according to the StatsSA (2011), Ekurhuleni’s total population is about 3.718 million of 

which black people are the dominant population group (see Appendix A). Seventy-one percent 

of the metropolitan's population is in the age group between 15 to 64 years. About 49.4 % of the 

population is employed while 19.9% is unemployed, and amongst those that are employed the 

average income is R29 400 per annum (StatsSA, 2011).  

The Metropolitan’s economy is dependent mainly on the manufacturing sector which is also the 

sector that produces employment for most residents. However, due to the decline of the mining 

sector and restructuring of industries, it has shaken the economy of the metropolitan area. This 

has been a great knock for Ekurhuleni and has resulted in challenges such as the loss of jobs 

(Todes et al., 2010). However, developments (such as retail development, public sector 

investment, and aerotropolis) have been emerging to re-attract investments and bring income 

into the metropolitan area (Rogerson, 2018; Todes et al., 2010). The economic fluctuations and 

spatial variations of development within the metropolitan area make the study site suitable for 

this research study. 

3.3 Data Acquisition 

The data used was collected based on the three analytical dimensions namely, Jobs and local 

economy, Education, Community health and civic life as indicated by the study of Wilson and 

Greenlee (2016). The study followed the model of Wilson and Greenlee (2016) where the three 

dimensions are used for measuring the opportunity index. Measuring these dimensions is 

important even in the context of South Africa because they are key determinants of a 

neighbourhood's outcome. For instance, measuring the economic standards of the 

neighbourhood could enable officials to make informed decisions with regards to strengthening 

the economy or developing the area. 

 

The main places have been used as the unit of study. The main place is level five in the 

geographical area hierarchy structure of StatsSA. It is above the sub-place and two levels below 

the metropolitan area (Frith,2011). All the data were obtained at the main place level, therefore 

there was no need to aggregate it. The main places that form part of Ekurhuleni metropolitan 



19 
 

area are as follows; Alberton, Bapsfontein, Benoni, Boksburg, Brakpan, Breswol, Centurion, Chief 

A Luthuli Park, Clayville, Daveyton, Duduza, Dukathole, Edenvale, Ekurhuleni NU, Etwatwa, 

Geluksdal, Germiston, Harry Gwala, Kanana, Katlehong, Kempton Park, Kwa-Thema, Langaville, 

Lindelani Village, Nigel, Springs, Tembisa, Thinasonke, Thokoza, Tsakane, Tweefontein, 

Vosloorus, and Wattville.  

 

Data for the local economy and education dimensions were acquired from the Census 2011 of 

Statistics South Africa. The community health and civic life data were acquired from the 

Department of Health (2011), Pick n Pay (2015) and Spar (2015). There was no available data close 

to the year the census data was collected, therefore we had to use data that was available. The 

indicators for Jobs and local economy dimension include the unemployment rate, household 

median income and poverty rate. The unemployment rate refers to the percentage of people 

that are not employed. This indicator helps to understand the level of unemployment.  The 

median income is the income that is in the middle of the overall income distribution, while the 

poverty rate is the percentage of people with the income that is below the poverty line. This 

dimension shows the economic and job standard of the neighbourhood (Wilson and Greenlee, 

2016). The education dimension includes indicators like pre-school enrolment, dropout rates, 

and post-secondary education attainment.  

 

The pre-school enrolment is all the children between 3 to 4 years-old that are enrolled at school. 

The dropout rate refers to the number of residents between the ages of 16 and 19 who are 

neither enrolled in school nor high school graduates, divided by the number of individuals aged 

16 to 19. The pre-school enrolment indicates the number of children that attending pre-school 

education while the dropout rate indicates the percentage of residents that have dropped out 

from school. Post-secondary education attainment is the percentage of people with post-

secondary education qualifications (Wilson and Greenlee, 2016). The education indicators 

indicate the level of education of a particular area. Health institutions, grocery stores, and youth 

inclusion are indicators of the community health and civic life dimension. Health institutions are 

the number of clinics per 100,000 people, while grocery stores include all the grocery stores per 



20 
 

10,000 people. Youth economic and academic inclusion refers to the residents between the ages 

of 16 and 19 who are enrolled in the educational institution and also not working (Wilson and 

Greenlee, 2016). These are indicators that promote a sense of belonging and provide networking 

platforms amongst the members of the neighbourhood (Rakodi, 2002; King, 2011). Table 1, 

shows the summary of datasets used and their sources.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.4 Data analysis 

Microsoft Excel was used to calculate the dimensions, and indices, while two software 

programmes, namely ArcMap 10.5.1 and R-studio 8.10, were interchangeably used to analyse 

the data. 

3.4.1 Data pre-processing 
 

Table 1:datasets and their sources 

 



21 
 

3.4.1.1 StatsSA attribute data 

The StatsSA data was provided in a spreadsheet format with different indicators. The data was 

cleaned by removing all variables/indicators not relevant to this project. The database was 

subsequently joined with the administrative boundary shapefile of Ekurhuleni’s main places for 

analysis. 

3.4.1.2 Point attribute data 

Data for the health facilities and grocery stores were acquired in shapefiles format (particularly 

points). For analysis, new fields were created with facilities per main place to calculate it based 

on the population it serves. i.e. health facilities per 100,000 residents and retail stores per 10,000 

residents. The metadata for all the point data is included in Appendix B. 

3.4.2 Descriptive statistics on the raw data  
 

The statistical description is a principal component of the exploratory study that provides 

summaries of the raw data, and also shows relationships between the variables (Kaushik and 

Mathur, 2014; Sharma, 2019). This analytical method can be presented numerically or graphically 

(Kaushik and Mathur, 2014). The numeric representations have been used to illustrate the 

distribution, the central tendency, and the variation in the data. 

3.4.3 Creation of the Opportunity Index 
 

The opportunity index was created using the data collected for each dimension. The creation of 

the opportunity index helps achieve the first objective of this study. The method of calculating 

the opportunity index includes normalizing the indicators using the observed values of the 

indicator. The indicators were normalised by comparing the observed raw data with the 

minimum and maximum values as indicated by the following formula (Wilson and Greenlee, 

2016). 



22 
 

 

 

 

 

The normalized indicators were averaged (by adding the rescaled values and divide by the total 

number of indicators for each main place) to create the scores for each dimension. Then the 

scores of all the dimensions were averaged (adding all the dimension scores and divide by the 

total number of dimensions for each main place) to calculate the opportunity index.   

3.4.4 Choropleth Mapping 
 

The Choropleth map is one of the ESDA tools which are used to visualise the spatial attributes 

(Leonowicz, 2006). The maps were used to show the spatial distribution of opportunities and 

population within the metropolitan area, and also illustrate the relationship between the 

opportunity and neighbourhoods. The second objective of the study is achieved in this section of 

the methodology. 

3.4.5 The Neighbours and spatial weights matrix 

Spatial analysis is based on the assumption that close phenomena are more similar than those 

that are farther apart (Foster and Evans, 2008). Therefore, it is crucial to investigate the presence 

of spatial patterns before data analysis. Spatial neighbours and weights determine the spatial 

relations between the spatial units (Chen, 2012; Cheng et al., 2014; Shen,1994). Distance-based 

weight is appropriate since it determines the relations between spatial units based on a specified 

distance. Distance-based weights could be measured using the distance two spatial units, it is 

also called Euclidean distance. It can also be measured using the distance between the centroids 

of polygons which is also called Great circle distance (Waller and Gotway, 2004; Bivand et al., 

2008). R studio was used to define the spatial neighbours and weights in which main places were 

used as the units of analysis. 

3.4.6 Spatial autocorrelation 
 

Equation 1: 



23 
 

There are two spatial autocorrelation methods (Global and local Moran). However, for this 

research study, only the local Moran was used to examine the spatial relations between the main 

places and the opportunities. Local Moran was chosen because the study site is more local. 

Local Moran’s I   

The local Moran is a local indicator of spatial autocorrelation which quantifies the degree of 

relations between smaller scales which may be overlooked by the Global measure (Moran’s I) 

(Wilson and Greenlee, 2016; Anselin, 1996). It also identifies clusters of a high and low distributed 

residuals whereby negative values indicate spatial outliers, and the positive values indicate 

clusters (Tiefelsdorf, and Boots,1997; Wilson and Greenlee, 2016; Anselin, 1996). The local 

Moran was used to measure the spatial autocorrelation between the main places.  

 

3.4.7 Hot spot and cold spot analysis 
 

The hot spot analysis is an extension of the local Moran analysis, it measures the spatial 

significance of the areas such that the area forms part of clusters and outliers. Clusters of high 

values indicated by High-High and low values indicated by Low-Low. The outliers are indicated by 

High-Low and Low-High (Wilson and Greenlee,2016). The analysis was adopted to examine areas 

with high (hot) and low (cold) clusters of opportunity. This part of the spatial analysis (including 

spatial autocorrelation) help solve the third objective of this study.  

3.5. Limitations  
 

The first limitation of this research approach is that it is mainly dependent on census data which 

was last updated in 2011. Since 2011, changes have occurred spatially, socially, and economically 

within the metropolitan area. This could have impacted the results in the case where the 

measured variables may have changed. Another limitation is that census data, and other data 

that have been used have different collection years. For instance, some datasets were collected 

in 2015 (which is four years) after the StatsSA data was collected. If all the data was collected at 

the same time it would be easier to compare it. However, the data is relevant and useful for 

examining the geography of opportunity in the metropolitan area. Lastly, the limitation of the 

https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Tiefelsdorf%2C+Michael
https://onlinelibrary.wiley.com/action/doSearch?ContribAuthorStored=Boots%2C+Barry


24 
 

methodology lies in the fact that exploratory analysis does not have a formal procedure of how 

the analysis should be structured; therefore, it is difficult to determine how far the analysis 

should be performed. 

  



25 
 

CHAPTER 4: RESULTS 

4.1 Introduction 
The use of different exploratory angles is essential for a clear understanding of the spatial 

dimension of opportunity in the Ekurhuleni metropolitan. The result first presents the statistical 

description of the indicators data that were used to create the opportunity index. The second 

section outlines the results of the analytical steps indicated in the methodology section. These 

analytical steps include the creation of the opportunity index, choropleth mapping of the 

opportunity index, detecting the spatial autocorrelation, hotspots, and cold spots. This section 

unravels the findings that answer the objectives of the research. 

4.2 Descriptive statistics 
 

Numerical and graphical representations of the data have been used. Table 2, shows the 

descriptive statistics for the jobs, and local economy dimension which comprises the median 

household income, poverty rate, and unemployment rate indicators. It is clear from the table 

that, there is less deviation from the mean for the poverty rate indicated by the low value of the 

standard deviation, while the high value of the unemployment rate and median household 

income’s standard deviations show a notable variance from the mean. The difference between 

the minimum, maximum, and median indicate that the data for the unemployment rate indicator 

is skewed to the left. While the poverty rate and median household income indicator are skewed 

to the right.  

The table also indicates that some of the main places have the unemployment rate of 2% and it 

goes up to 46% for some of the main places as indicated by minimum and maximum values. The 

poverty rate goes between 0 and 12% while the median income ranges from 3 to 16 957 for some 

of the main places. Main places with a high unemployment rate also have high poverty rates with 

several residents earning the median income.  

 

 

 



26 
 

 

 

 

 

 

 

 

 

 

Table 3, shows that as compared to jobs and local economic indicators, all indicators of the 

community health and civic life show less deviation from the mean as indicated by the low value 

of the standard deviations. The difference between the minimum, median, and maximum values 

for Youth inclusion and health institutions per population indicators show that they are heavily 

skewed to the right while the grocery store per population is slightly skewed to the right. The 

most popular observation for retail outlets and health institutions per population is 0 indicating 

a lack of grocery stores and health institutions in most of the main places.  

 

 

 

 

 

 

 

The standard deviations of all the Education indicators are low (see Table 4) showing less 

deviation of the data from the mean. The minimum, median, and maximum values indicate the 

right skewness in the data. The higher education attainment indicator has 0 as a popular 

Table 4:Descriptive Statistics for Jobs and local economy indicators 

 

Table 7:Descriptive Statistics for Community health and Civic life indicators 

 



27 
 

observation indicating the dominance of residents that do not have tertiary educational 

qualifications.   

 

 

 

 

 

 

 

 

 

 

4.3 Opportunity Index 
 

The opportunity index scores show variation amongst the main places, some have high while 

others have low scores (see opportunity index Table 5). The scores range from 2 to 49. The two 

main places that have the lowest scores of below 20 are Tweefontein and Kanana. Tweefontein 

has zero scores for education and civic life dimensions which resulted in the lowest overall 

opportunity score. Unlike Tweefontein, Kanana has comparable scores for all the dimensions 

which also boosted the overall opportunity. The main places that have the highest opportunity 

score of above 46 are Germiston and Nigel. Germiston scored above 50 in both the local economy 

and civic life dimensions while Nigel scored above 50 in Civic life dimension only, which resulted 

in the highest opportunity scores.  Generally, most of the main places with high opportunity are 

developed towns while a few are townships. Table 5 shows the numerical representation of the 

opportunity indexes and dimension scores. 

 

Table 10:Descriptive Statistics for Education indicators 



28 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The dimension scores also show variations amongst the main places. Some main places scored 

higher in all the dimensions which resulted in a high score for the opportunity index. Other main 

places scored higher and lower in other dimensions which led to lower scores for the overall 

opportunity. The descriptive statistics were also presented to relate the dimensions and the 

opportunity as shown in Table 6. It is apparent from the table that Jobs and the local economy 

dimension scored higher on average with a score of 35.5 followed by education with a score of 

Table 13:Opportunity Index 



29 
 

31.5 while community health and civic life scored the least score of 28.8. Jobs and local economy 

also scored higher than the opportunity index on average. The standard deviations of the 

dimensions also follow a similar trend as that of the average scores. Jobs and local economy have 

the highest standard deviation with the opportunity index having the lowest. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

According to Wilson and Greenlee (2016), banking institutions and households spending on 

rental housing were also part of the jobs and local economy indicators. However, due to the 

scarcity of the data these indicators were not included. The rental expenditure data was only 

acquired at the metropolitan level (see table 7). While it does not give detailed information about 

the rental expenditure per main places, it gives a general overview of the metropolitan’s rental 

expenditure. Ekurhuleni is the municipality that has the lowest number of households renting 

Table 16:Descriptive statistics of the opportunity index 

 



30 
 

and spending their income on rental houses followed by Tshwane, while the city of Johannesburg 

has the highest number of households spending their income on rental housing. 

 

 

 

 

 

 

 

 

 

 

 

4.4 Mapping the distribution of opportunity 
 

The choropleth mapping includes the maps of the opportunity index and the individual 

dimension. The choropleth map of the opportunity index is also overlaid with the total population 

to relate the distribution of opportunity to the total population. The number of population for 

each main place is indicated in table 5, while the population groups per main place are included 

in Appendix C. 

4.4.1 Opportunity Index choropleth map 
 

The opportunity index choropleth map shown in figure 5, was created using the opportunity 

index scores indicated in table 5. The map presents the distribution of opportunity in space. The 

spatial manifestation of the opportunity shows that there are variations in the opportunity 

distribution within Ekurhuleni. The opportunity distribution ranges from a very high to a very low. 

Germiston, Boksburg, Nigel, and Dukathole are the four main places that have the highest 

opportunity scores. Amongst these four main places, three are developed towns while one is a 

Table 19:Rental expenditure statistics 



31 
 

township. It can also be observed from figure 2 that the distribution of high opportunity is 

predominantly in areas known as towns while low and very low distribution of opportunity is 

predominant in areas known as townships. As also indicated by table 5 that the main place with 

the highest opportunity index score is Germiston (developed town) and the lowest is Thinansoke. 

 

The findings show that there is no clear link between the distribution of opportunity and the 

area’s total population. For example, some main places (see figure 3) like Tembisa and Katlehong 

have a high distribution of opportunities with a high number of people staying in them. While 

main places like Nigel (developed town) do not have many people but have a very high 

opportunity distribution. There are also main places like Duduza, Vosloorus, Tsakane, and Kwa-

Thema that have a low distribution of opportunity with a high number of people staying in them. 

Figure 3: Opportunity index choropleth map (Author, 2020) 

 



32 
 

These main places are located at the peripheral locations, and the towns near them are well 

serviced. For instance, townships like Tsakane have a poor opportunity with a high population, 

however, it is located next to Nigel town which is well-developed and having advanced 

opportunities. Generally, the metropolitan's high opportunity is concentrated in main places 

considered to be well-serviced towns while low opportunity population is concentrated in 

townships. Figure 3, shows the opportunity index per total population of each main place. One 

dot on the map indicates 2,500 people. 

 

 

 

 

 

  

Figure 4: Opportunity index choropleth with total population distribution(Author,2020) 

 



33 
 

The physical appearance of space in some of the main places do resemble the level of opportunity 

they have while others do not. Main places (particularly towns) that have good opportunity 

distribution are also physically appealing, while some have different characters of areas that have 

a good opportunity. Images have been captured to illustrate the physical appearance of some of 

these main places. It is expected that the main places that have a high opportunity score should 

have aesthetically pleasing spaces, however, it is not the case for some main places.  Boksburg is 

one of the developed towns characterised by high opportunity distribution. Figures 4 and 5 show 

the physical appearance of some parts of Boksburg. The town has unpolluted streets and well-

planned buildings. Land use is also integrated (different land use close to each other) to enables 

access to different opportunities within proximity. There are also street lights to enhance safety 

and security in the area (see Figure 4).     

 

 

 

 

 

 

 

 

 

 

 

 

As shown in figure 5, the main place has recreational activities that act as the tourism centre that 

attracts the main place's income and for entertainment. Figure 5, shows the Wild water 

recreational area in the Boksburg main place. The area is the tourism site in which both local and 

international residents spend their summer days. It is clear from the figure that the area is well 

serviced and attract people of different races. 

Figure 5: Image showing some parts of Boksburg town( Van Zyl, 2013) 

 



34 
 

 

 

 

 

 

 

 

 

 

 

 Compared to Boksburg, Kempton Park main place has spaces that are visually appealing and 

indicate that the area is well-developed. Figure 6, shows the OR Tambo International Airport 

which is located in Kempton Park (City of Ekurhuleni, 2019). The airport is the transport hub for 

international and local travellers and is declared as the metropolitan area’s aerotropolis (City of 

Ekurhuleni, 2019).  

 

 

 

 

 

 

 

 

Figure 6: Image showing the recreational area in Boksburg( Wang, 2018) 

 

Figure 7: Image showing OR Tambo International airport( TimesLIVE, 2017) 

 



35 
 

 

The townships like Katlehong and Tembisa were classified as a high opportunities distribution 

area with a high number of people staying in them. However, unlike Boksburg and other 

developed towns, these townships do not visually look good or resemble an area with high 

opportunity. Figure 6, shows an image of Katlehong township. The township is still deprived; 

some parts of the area have informal dwellings that are built without a proper plan. There are no 

proper streets or street names that could be useful for navigating through the area. The trees in 

figure 10 show that the settlement has long been in existence, hence many people are staying in 

it. Despite the high number of people, the residents still use communal buckets toilets (see Figure 

7).  

 

 

 

 

 

 

 

 

 

 

 

Dukathole is another main place that was classified as a very high opportunity area. This main 

place is bounded by both Germiston and Boksburg (both developed towns). This main place 

consists of both informal and formal dwellings. Structurally, the area is not properly planned, and 

according to the statistical data, the main place has high unemployment and poverty rate. It is 

clear from figure 11 that the area is underprivileged.  Figure 8, shows the location of the school 

and informal structures along Peterson road. It is apparent from the figure that unlike Katlegong 

Figure 8: The physical appearance of some parts of Katlehong  township (Janine, 2018 ) 

 



36 
 

township, Dukathole is electrified, and has formal streets with street names, but they are not in 

good condition. Judging from the physical appearance of the area, one could see that the area 

does not indicate good opportunity as indicated by the opportunity index results. 

 

 

 

 

 

 

 

 

 

Among the main places that were classified as low opportunity areas, is Thokoza township. The 

area does resemble areas of poor opportunity as indicated by Figure 9. It can be observed from 

the figure that Thokoza also has a cluster of shacks.  

 

 

 

 

 

 

 

 

Figure 9: Location of school and shacks along Peterson road in Dukathole (Street view, 
2017) 

Figure 10: Cluster of shacks and informal trading in Thokoza township (Khumalo, 2014) 

 



37 
 

4.4.2 Jobs and local economy choropleth 
 

The individual dimension gives a clear sense of what opportunity index comprises. Therefore, it 

is crucial to understand the distribution of each dimension in space. Figure 10, indicates the 

distribution of jobs and the local economy’s dimension distribution. It is clear from the figure that 

the distribution of jobs and local economy differs from the overall economy. Very high 

distribution is noticed in three, main places namely Katlehong, Etwatwa, and Tembisa. Two of 

the three main places (Katlehong and Tembisa) also have higher opportunity distribution while 

Etwatwa has a low distribution. Unexpectedly, Kempton Park is characterised by the lowest 

distribution on jobs and local economy. This main place has a higher score of opportunity 

distribution and it is also known as one of the developed towns. Although it has a higher 

distribution of overall opportunity, the standard of its local economy is proven to be vulnerable 

(see figure10).   

 

 

 

 

 

 

 

 

 

 

 

 

Figure 11: Choropleth map for jobs and local economy dimension(Author, 2020) 

 



38 
 

The other main places that have a low distribution of jobs and local economy are Edenvale 

Alberton and Clayville. Among the three, Clayville and Edenvale also have low opportunity 

distribution while Alberton has a high distribution. As compared to Kempton Park, Edenvale and 

Alberton are known to be the developed towns. Therefore, it was unexpected that they could 

have a low distribution of the local economy. 

  

4.4.3 Education 
 

The education dimension choropleth shows a different trend to that of jobs, the local economy, 

and that of the opportunity index (see figure 11). Four developed towns have a very high 

distribution of education dimension. Two (Germiston and Boksburg) of these towns also have the 

highest distribution of overall opportunity while the other two have higher distribution. All the 

main places that have the lowest distribution range of education are townships located in the 

peripheral areas. 

Kempton Park has a very high distribution of education dimension which is different from that of 

jobs and local economy dimension. Images were used to illustrate the quality of schools in some 

main places.  Figure 12 shows one of the schools in Boksburg, the school is well maintained, clean, 

and liveable for children. According to Powell et al (2007), children in well-advanced schools tend 

to perform better than those in poor schools.  

 

 

 

 

 

 

 



39 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 12: Choropleth map for Education(Author, 2020) 

 

Figure 13: School in Boksburg (Grove, 2010) 

 



40 
 

4.4.4 Community health and civic life 
 

The community health and civic life dimension scores indicated a completely different 

distribution compared to other dimensions, and this opportunity index as shown in figure 13. The 

towns that are characterised by a very high distribution of opportunity as indicated in figure 5, 

are all associated with a low distribution of community health and civic life opportunities. Most 

of the places that have a high distribution of community health and civic life opportunities are 

the areas that also have a very low overall opportunity. Brakpan is the main place with a high 

distribution of high community health and civic life but having the lowest opportunity score. This 

indicates that although the economy and the education standard of the area is poor, there is a 

strong interaction between the residents. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
Figure 14: Choropleth map for community health and civic life dimension(Author, 2020) 

 



41 
 

Access to grocery stores is one of the community health and civic life indicators (Wilson and 

Greenlee, 2016). The availability of the grocery stores that are supporting a specific number of 

people improves the level of opportunity in the area. Boksburg town has a variety of stores that 

provide adequate and healthy goods. These stores also provide goods even for people in other 

main places (Evans,2014). Unlike in Boksburg, residents in townships like Duduza depend on 

spaza shops (see figure 14) for the supply of goods on their daily bases. These informal stores do 

not sell all the required goods since they are small businesses ran by fellow residents. Therefore, 

residents have to travel to other areas to buy goods. Moreover, the quality of these foods can 

also be questioned considering the state the spaza shops, and the time that the goods spend 

without being sold. Some goods are sold by street vendors (see figure 15).  These goods are 

exposed to the sun and could easily be infected by bacteria. Therefore, residents might have 

access to grocery stores in the township, but the quality of food offered can be questionable. 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 15: Grocery stores in Duduza township(Baloyi, 2013) 

Figure 16: Street vendors in Thokoza township(Andrews, 2009) 

 



42 
 

4.6 Examining the spatial pattern of opportunity 
 

4.6.1 Local Moran 
 

The local Moran results show the spatial autocorrelation and significant difference between the 

main places’ opportunity. It can be observed from the map in figure 16, that there are spatial 

clusters (indicated by the positive values) and spatial outliers (indicated by negative values). The 

areas that have a clustering of opportunities are Boksburg, Germiston, Kempton Park, Benoni, 

Springs, Nigel, Brakpan, Alberton, Katlehong, Etwatwa and Tembisa (showed by positive values), 

while the rest are outliers (showed by negative values). However, it is not clear whether the 

clusters are (high) or cold (low), hence the need to perform the hot spot analysis. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 17: Local Moran statistic map(Author, 2020) 



43 
 

4.6.2 Hot spot and cold spot clusters 
 

Since the local Moran statistics only indicated the clusters without indicating whether they are 

hot, and cold spots, figure 17, categorises the clusters as high and low. It can be seen from the 

figure that Boksburg, Germiston, and Kempton Park are the only main places that have a high 

cluster (hot spots) of opportunities while Tweefontein has low cluster (cold spots) of 

opportunities while the rest was classified as not significant. 

  

Figure 18: Hot and cold spot map(Author, 2020) 



44 
 

4.7 Conclusion 

The raw data of the indicators has given an understanding of where the opportunity indices came 

from. They indicated why some main places scored higher in their dimensions which also boosted 

the overall score of the opportunity index. The mapped opportunity index showed that the 

distribution of opportunity differs between the privileged and underprivileged communities. It 

was also observed that there is no clear relationship between the distribution of population and 

the distribution of opportunity as there are townships that have high population distribution but 

different opportunity distribution. Some main places were classified as high opportunity areas 

but they are not physically pleasing.  

The individual dimension has a different distribution trend in comparison to the coverall 

opportunity index. Some main places have a low distribution of the overall opportunity while 

having a high distribution of dimensions. This indicates that the area might have a low 

opportunity score, but performing well in some of the other dimensions. 

 

 

 

 

 

 

 

 

 

 



45 
 

CHAPTER 5: DISCUSSION OF FINDINGS 

5.1 Introduction 

The findings have generally proven the uneven distribution of opportunity between the 

privileged and vulnerable areas of the metropolitan area. This chapter elaborates more on these 

variations and interprets their implications. The first section of the chapter will reflect on the 

implications of the main results while the second segment will present the reflection of the 

findings relative to the arguments presented by the literature. The scholars' arguments will be 

drawn from the themes such as the hypothesis of spatial mismatch, neighbourhood segregation, 

and livelihoods as presented in the literature review. The overall reflections from this section will 

direct the concluding remarks of the research. 

5.2 Reflection on the significance of the main findings 
 

The main results give a clear perspective of the spatial dimension of the overall opportunity and 

individual dimensions. Therefore, the purpose of reflecting on these findings is to present their 

implications for this study. 

5.2.1 Descriptive statistics 
 

The statistical descriptions have indicated the dominance of 0 observation in some opportunity 

indicators. Zero observation is a bad indication for some, and good indication for other indicators. 

For instance, the dominance of 0 observation in the Higher education attainment indicator 

implies that most of the residents within the metropolitan area do not have an associate or higher 

degree. A few main places have notable high unemployment and poverty rate and these two 

indicators have a positive relation to each other. Implying that the main places that were found 

to have a high level of unemployment also have high rates of poverty. There is a direct link 

between poverty and unemployment i.e. high rates of unemployment can result in induced 

poverty intensity (Cloete, 2015). One of the causes of poverty is that the country’s economic 

systems promote self-reinforcement of other areas while others remain poor (David et al., 2018). 

The country provides free services such as social grants to combat these challenges, however, 



46 
 

poverty persists in the country (David et al., 2018). The results showed that most of the main 

places have a high percentage of the unemployment rate and according to Cloete (2015), 

unemployment has racial and sexual implications in South Africa. Despite the skills and 

educational qualifications, black South Africans (mainly women) are unemployed while the 

whites tend to be employed. 

The indicators have also shown the dominance of the high school dropout rate and the shortage 

of higher education attainment in most of the main places. This reveals that the standard of 

education is a social issue in some parts of the metropolitan area. According to Gregorio and Lee 

(2010), education has a strong relation to the distribution of income within the area. Therefore, 

the attainment of higher associate degrees can impact the distribution of income within the area. 

The level of education explains why the income is unequally distributed within the municipality. 

The poor standard of education also explains why some of the main places have high 

unemployment rate because the more residents drop out from high school, the more they will 

not attain tertiary qualifications, and not acquire necessary skills to be competent for 

employment (the author is aware that poor standard of education is not the only cause of 

unemployment but is very significant).   

The result also indicated that most of the main places do not have many health institutions and 

grocery stores. This is an indication that most of the population compete for the existing facilities 

which increase infrastructural pressure to the institutions or induce distance travelled by 

residents to access these facilities. There is evidence that poor geographic accessibility of 

facilities results in the loss of lives, therefore, it is needed to locate health facilities at an optimal 

location to accommodate many people (Rahman and Smith, 2000). The World Health 

Organization (WHO) has mandated the nations to ensure that health facilities should be 

affordable and accessible to the citizens at the minimum distance of 5 km, this has not been a 

case in many countries. People are still forced to travel long distances to access health services 

(McLaren et al., 2013; Kemboi and Waithak, 2013). Other than distance frictions, health facilities 

are exposed to the pressure of servicing the population beyond their capacity. Therefore, a 

shortage of health facilities in most of the Ekurhuleni’s main places indicates the residents are 

exposed to the challenges mentioned above. 



47 
 

5.2.2 Opportunity  
 

The presented opportunity index and the choropleth indicate the variation between the 

opportunity of the main places. Most of the main places that have very high scores are those that 

are known as the developed towns (Germiston and Boksburg) and one township (Dukathole) 

which is adjacent to the same developed towns. Surprisingly, the township that was classified as 

a very high opportunity area has a visual appearance that does not resemble an area with high 

opportunity. This means that the opportunity index does not consider appearance but focuses 

on a numeric analysis dependent on the dimensions applied. Therefore, it could have argued that 

the location of Dukathole with the developed towns might have influenced its opportunity score. 

Because there are no physical boundaries that separate main places, therefore some could be 

subjected to the spillover effects. This implies that the development of one area could result in a 

negative or positive spillover effect on the neighbouring areas. Thus, it is worth stating that 

Dukathole township is the product of the positive spillover effect from both Boksburg and 

Germiston.  

There are main places that scored lower and are mainly townships, most of these main places 

have low opportunity scores and many people staying in them. Ravallion (2001) named situations 

such as this ‘urbanisation of poverty’. Implying that the influx of people in a poor urban 

environment increases the level of poverty. This could be as a result of the high competition for 

the livelihood resources and infrastructural pressure. One could also argue that although these 

townships are poor, people choose to locate in them because of the affordable land and flexibility 

of building dwelling structures (particularly shacks). These are the poor townships mainly located 

at the peripheral locations or next to developed towns. One reason that induces their 

vulnerability is the self-reinforcement of the developed towns and the negative spillover effects. 

This implies that the developed towns continue to grow and as they grow the surrounding areas 

become poorer.  

The distribution of the opportunity does not follow the same trend as that of the individual 

dimensions. Areas that have a high distribution of opportunity also have a low score in some of 

the dimensions. With this, one could argue that a high distribution of opportunity does not 



48 
 

necessarily mean that all the dimensions are well presented in that particular area. Thus, 

indicating that the area can still be classified as the cluster of opportunity, although their level of 

education or standard of the economy is low. This also shows a limitation of classifying 

communities as clusters of opportunity because it is not an informed conclusion to refer to a 

community as a cluster of high opportunity while, for instance, people are economically 

struggling. 

5.2.3 Local Moran 

The local Moran showed the spatial autocorrelation between the main places. It was also 

observed from the Moran statistics map that, there are clusters of opportunity in some of the 

areas. The hotspot map indicated the High-High cluster and the outliers. The High-High indicates 

the clustering of high opportunity (also known as a hotspot). The areas with high-high clusters of 

opportunity are developed towns such as Germiston, Boksburg, and Kempton Park. Although 

some townships have a high distribution of opportunity as indicated by the choropleth maps, 

they are not classified as high-high clusters. This supports the statement made previously that 

although the townships are classified as areas of high opportunity distribution, they might still be 

socially, spatially, and economically vulnerable. 

5.3 Reflections on the findings in relation to the literature 

The literature was centred on themes such as the hypothesis of the spatial mismatch, 

neighbourhood segregation, and livelihoods. Therefore, this section will reveal how the research 

results could help to fill in the gaps identified in the literature review.  

5.3.1 The hypothesis of spatial mismatch theory 

The spatial mismatch has been discussed as the physical separation between the job 

opportunities and the residential area (Howell-Moroney, 2005; Naudé, 2008; Gobillon et al., 

2007). They further alluded that this happens because most of the economic activities are 

concentrated in developed areas while most people reside in areas with poor opportunities away 

from these economic activities. In the case where the townships are located close to economic 

activities, the residents do not have the necessary skills to participate in the economic activities 

offered (Howell-Moroney, 2005; Naudé, 2008). Another cause of the spatial mismatch was found 



49 
 

to be the housing markets whereby the houses provided next to the economic centres caters 

only the affluent group (Naudé, 2008; Powell et al., 2007). 

Ekurhuleni metropolitan area has proven to have a high distribution of jobs, and the local 

economy in few towns and few townships. The availability of jobs and economic activities enable 

residents residing in or near these places to have access to these opportunities at proximity. 

However, this is pertinent to residents that have relevant skills needed for the jobs provided. As 

for those that do not qualify or staying in distant areas will have to commute to access these 

opportunities. Scholars like Naudé (2008) argued that in some contexts spatial mismatch is a 

resemblance of what they termed ‘racial mismatch’. This phrase explains the lack of job 

opportunities where black people can be hired. Most of the townships in Ekurhuleni are 

dominated by black people, and clusters of low opportunity. Thus indicating that ‘racial 

mismatch’ could describe the spatial mismatch in Ekurhuleni metropolitan area.  

 

5.3.2 Neighbourhood segregation 

One of the apartheid's legacies in South Africa was to separate people based on race, class, and 

income. Black people have been pushed towards the peripheral townships away from economic 

activities (Mahajan, 2014; Peberdy et al., 2017). According to Powell et al (2007) the influence of 

land use regulations, the spatial sprawl also causes the isolation of communities. These 

marginalised townships are characterised by livelihood opportunities (Peberdy et al., 2017; 

Naudé, 2008; Powell et al., 2007). Townships like Duduza and Thokoza have the characteristics 

of those that are marginalised: poor opportunity, poor local economy, education, and community 

health opportunities. These neighbourhoods struggle to provide the necessary resources to 

sustain their people. Social issues such as unemployment rate, dropout rate, poverty rate, etc. 

are also persistent in these areas, thus, affect people’s outcomes. To make social issues worse, it