ResearchSpace :: Browsing by Author "Holloway, Jennifer P"

Browsing by Author "Holloway, Jennifer P"

Now showing 1 - 20 of 21

The 2011 municipal elections in South Africa and new trends since the 2009 national elections
(Operations Society of South Africa, 2011-09) Greben, JM; Elphinstone, CD; Holloway, Jennifer P
The CSIR has been involved in South African election night predictions since 1999 using a cluster prediction model based on the segmentation of the electorate according to voting behavior. In this paper these clusters are exploited in another way. Different clusters are related to different demographic groups, and an analysis is made how these different groups change their affiliation between subsequent elections. The changes in affiliation are determined by calculating a trend matrix, a new tool in elections that was introduced by one of the authors a few years ago. By comparing trend matrices between municipal (2006, 2011) and national elections (2004 and 2009) one can establish whether the observed trends are incidental or have a more generic character. It is felt that a better understanding of the voter behavior through such analyses can enhance the value of elections and thereby promote democracy.
2014 Election forecast - a post-election analysis
(Operations Research Society of South Africa (ORSSA), 2017-06) Ittmann, HW; Holloway, Jennifer P; Dudeni-Tlhone, Nontembeko
General elections are held every five years in South Africa. During the 12 to 24 hour period after the close of the voting booths, the expected final results are of huge interest to the electorate and politicians. In the past, the Council for Scientific and Industrial Research (CSIR) has developed an election forecasting model in order to provide the media and political analysts with forecasts of the final results during this period of peak interest. In formulating this model, which forecasts the election results as the results from voting districts (VDs) become available, some assumptions had to be made. In particular, assumptions were made about the clustering of previous voting patterns as well as the order in which VD results are released. This election forecasting model had been used successfully for a number of elections in the past and in these previous elections, with around 5% {10% of the results available, the predictions produced by the model were very close to the final outcome, particularly for the ANC, being the largest party. For the 2014 national election, however, the predictions, with close to 50% of the voting district results known (equivalent to an estimated 40% of the total votes), were still not accurate and varied by more than 1% for both the ANC and the EFF. This paper outlines a post-election analysis to determine the reasons for these discrepancies and how they relate directly to the model assumptions. The aim is to highlight how practical realities can affect the assumptions and consequently their impact on the forecasted results. Reference is made to previous election forecasts and the 2014 post-election analysis is presented.
A phenomenological methodology for wave detection in epidemics
(2025-12) Brettenny, W; Holloway, Jennifer P; Fabris-Rotelli, I; Dudeni-Tlhone, Nontembeko; Abdelatif, Nontembeko; Le Roux, Wouter J; Manjoo-Docrat, R; Debba, Pravesh
In both the management and modelling of epidemic outbreaks, the ability to determine the start of a wave of infections is of vital importance. Not only does this advantage the modelling of the outbreak, but, if done in real-time, can assist with a nation’s response to the disease. In this study, a bidirectional long-short-term-memory (Bi-LSTM) network is used to determine the start and end of the COVID-19 waves experienced in the district and metropolitan municipalities of Gauteng, South Africa, from 2020-2022 as well as the waves of the cholera outbreaks occurring in the Beira area of Mozambique between 1999 and 2005, in real-time. The problem of real-time scaling of the data prior to the first wave of an epidemic is addressed using globally available real-time information from first waves experienced in other countries and independent territories alongside the observed South African data. The use of the Bi-LSTM predicted starting dates is demonstrated for the second waves of COVID-19 infections experienced in Gauteng in 2020/21. Using the predicted starting dates, spatial-SEIR models are used to predict hospitalisations as a result of COVID-19 infections in each of the district and metropolitan municipalities of Gauteng. The fitted Bi-LSTM model demonstrates effectiveness in predicting the start and end dates of epidemic waves in real-time, allowing for pre-emptive disease modelling and predictions of spread. Moreover, it is shown that the use cases for the fitted model are not limited to COVID-19 studies, but can also be applied to other disease outbreaks that follow similar wave patterns.
Accessibility, affordability, and equity in long-term spatial planning: Perspectives from a developing country
(2022-05) Van Heerden, Quintin; Karsten, Carike; Holloway, Jennifer P; Petzer, Engela; Burger, Paul AD; Mans, Gerbrand G
City planners attempt to create more equitable spaces by providing and improving access to benefits of living in cities, especially for previously disadvantaged urban communities. To this extent, evidence-based decision making is required to adequately plan for and improve accessibility to several types of facilities. Accessibility studies in literature focus mostly on one element, which is sufficient when presenting methodological advancements, but it is limiting when providing decision support to city planners. This paper argues that these measurements should be expanded and there is a need for a nuanced view on accessibility for improved urban planning practices. Such a view is presented by simultaneously considering various categories of supply (employment, housing, transportation, health, education, police), multiple modes of transport (walking, private vehicle, numerous transit modes), two cost thresholds (distance-based and monetary cost), level of access (percentage of facilities that can be reached), while distinguishing between the socio-economic profiles of regions in the city on the demand side. This improves the understanding of affordability and equity in the study of accessibility. Furthermore, this paper expands two categories (education and housing) to explain the influence of capacity on accessibility and equity. Lastly, it couples a land-use model to some of the accessibility measures to show the usefulness of using such a combination in long-term spatial planning and what the effects will be without government intervention, again contributing to the understanding of, and planning for, more inclusive and equitable cities.
The applicability of the South African census 2011 data for evidence-based urban planning
(Statistics SA, 2016-06) Khuluse-Makhanya, Sibusisiwe; Dudeni-Tlhone, Nontembeko; Holloway, Jennifer P; Schmitz, P; Waldeck, Louis; Stein, A; Debba, Pravesh; Stylianides, Theo; Du Plessis, Pierre; Cooper, Antony K; Baloyi, Ethel
In urban planning it is important to understand settlements in terms of demographic, socio-economic, physical and political environmental characteristics. The objective is often to support municipalities, regional and national governments with long-term planning tools for the development of infrastructure, facilities and services. In a municipal context planning relies on spatially explicit estimates of the future demand for services, which depend largely on where households will live, where they will work and how they will commute using the transportation networks available to them. Given the complexity of large cities as a system, modelling and simulation tools are valuable for municipal decisions regarding the likes of environmental quality and safety in settlements and stimulation of economic productivity through provision of infrastructure. Models cannot capture such a complex system entirely; however they do allow integration of data indicative of factors shown in Figure 1 and their dynamics simplistically, for consumption by the relevant stakeholders. That is, they provide evidence upon which spatially equitable and sustainable planning of urban areas can be based.
An application of the Autoregressive Conditional Poisson (ACP) model
(SASA 2010, 2010-11) Holloway, Jennifer P; Haines, L; Leask, K
When modelling count data that comes in the form of a time series, the static Poisson regression and standard time series models are often not appropriate. A current study therefore involves the evaluation of several observation-driven and parameter-driven time series models for count data. In the observation-driven class of models, a fairly simple model is the Autoregressive Conditional Poisson (ACP) model. This presentation will describe the formulation of this model, together with the extension to the Double Autoregressive Conditional Poisson (DACP) model and also present some results of how these models compare to the static Poisson regression when modelling example data of cholera epidemics.
Clustering of housing and household patterns using 2011 population census
(2013-11) Dudeni-Tlhone, N; Holloway, Jennifer P; Khuluse, S; Koen, Renée
This study looked at a specific application of cluster analysis using the recently released population census 2011 data for the Ekurhuleni Metro in the Gauteng Province of South Africa. The main focus of the clustering was to distinguish housing and household patterns in order to create homogenous groups with similar demands for infrastructure, facilities and services. The k-means algorithm was specifically applied to groups of variables (factors) such as the dwelling types, conditions and location characteristics, socio-economic profiles, as well as demographic factors. These groups of clusters were later combined in a sequential manner to obtain a final set of meaningful clusters that could be used as inputs into an urban growth simulation tool.
Comparison of various modelling approaches applied to cholera case data
(2008-06) Van Den Bergh, F; Holloway, Jennifer P; Pienaar, M; Koen, Renée; Elphinstone, CD; Woodborne, S
The application of a methodology that proposes the use of spectral methods to inform the development of statistical forecasting models for cholera case data is explored in this paper. The seasonal behaviour of the target variable (cholera cases) is analysed using singular spectrum analysis followed by spectrum estimation using the maximum entropy method. This seasonal behaviour is compared to that of environmental variables (rainfall and temperature). The spectral analysis is re ned by means of a cross-wavelet technique, which is used to compute lead times for co-varying variables, and suggests transformations that enhance co-varying behaviour. Several statistical modelling techniques, including generalised linear models, ARIMA time series modelling, and dynamic regression are investigated for the purpose of developing a cholera cases forecast model fed by environmental variables. The analyses are demonstrated on data collected from Beira, Mozambique. Dynamic regression was found to be the preferred forecasting method for this data set
Could the outcome of the 2016 US elections have been predicted from past voting patterns?
(2017-07) Schmitz, P; Holloway, Jennifer P; Dudeni-Tlhone, Nontembeko; Ntlangu, Mbulelo B; Koen, Renée
In South Africa, a team of analysts has for some years been using statistical techniques to predict election outcomes during election nights in South Africa. The prediction method involves using statistical clusters based on past voting patterns to predict final election outcomes, using a small number of released vote counts. With the US presidential elections in November 2016 hitting the global media headlines during the time period directly after successful predictions were done for the South African elections, the team decided to investigate adapting their method to forecast the final outcome in the US elections. In particular, it was felt that the time zone differences between states would affect the time at which results are released and thereby provide a window of opportunity for doing election night prediction using only the early results from the eastern side of the US. Testing the method on the US presidential elections would have two advantages: it would determine whether the core methodology could be generalised, and whether it would work to include a stronger spatial element in the modelling, since the early results released would be spatially biased due to time zone differences. This paper presents a high-level view of the overall methodology and how it was adapted to predict the results of the US presidential elections. A discussion on the clustering of spatial units within the US is also provided and the spatial distribution of results together with the Electoral College prediction results from both A ‘test-run’ and the final 2016 presidential elections are given and analysed.
Economic scenarios and their impact on spatial transformation: A Johannesburg case study
(2025-06) Karsten, Carike; Holloway, Jennifer P; Mans, Gerbrand G; Beetge, Carla
This paper examines how different economic growth scenarios affect the success of Spatial Development Frameworks (SDFs) in shaping urban development in the Johannesburg Metropolitan Municipality. Using the CSIR’s urban growth model, the study compares two scenarios: low economic growth and recovering economic growth, to assess their impact on spatial outcomes such as informal settlement growth, urban sprawl, and housing density. The findings reveal that economic growth significantly influences the effectiveness of SDFs—higher economic growth leads to better land use management, reduced informality, and more successful densification in central areas, whereas low economic growth exacerbates informal housing and sprawl.
Exploring the usefulness of the INLA model in predicting levels of crime in the City of Johannesburg, South Africa
(2024) Coleman, Toshka; Mokilane, Paul M; Rangata, Mapitsi R; Holloway, Jennifer P; Botha, Nicolene MR; Koen, Renée; Dudeni-Tlhone, Nontembeko
Crime prediction serves as a valuable tool for deriving insightful information that can inform policy decisions at both operational and strategic tiers. This information can be used to identify high-crime areas, and optimise resource allocation and personnel management for crime prevention. Traditionally, techniques such as the Poisson model and regression analysis have been widely used for crime prediction. However, recent statistical advancements have introduced Integrated Nested Laplace Approximations (INLA) as a promising alternative for spatial and temporal data analysis. This study focuses on crime prediction using the INLA model. Specifically, the first-order autoregressive model under the INLA modelling framework is employed on longitudinal data for crime predictions in different regions of the City of Johannesburg, South Africa. The model parameters and hyperparameters considering space and time are estimated through the INLA model. In this work, the suitability and performance of the INLA model for crime prediction is assessed, which effectively captures spatial and temporal patterns. This study contributes to research by first introducing a novel approach for South African crime prediction. Secondly, it develops a model using no demographic information other than clustering attributes as an exogenous variable. Thirdly, it quantifies prediction uncertainty. Finally, it addresses data scarcity through demonstrating how INLA can provide reliable crime predictions, where conventional methods are limited. Based on our findings, the INLA model ranked areas by crime levels, obtaining a 29.3% Mean Absolute Percentage Error (MAPE) and 0.8 value for crime predictions. These findings and contributions presents the potential of INLA in advancing evidence-based decision-making for crime prevention.
From SA to the USA: Election forecasting
(AJOL, 2018) Holloway, Jennifer P; Ittmann, HW; Dudeni-Tlhone, Nontembeko; Schmitz, P
Elections draw enormous interest worldwide, especially if these involve major countries, and there is much speculation in the media as to possible outcomes from these elections. In many of these recent elections, such as the UK and USA, however, forecasts from market surveys, electoral polls, scientific forecasting models and even exit polls, obtained from voters as they leave the voting stations, failed to predict the correct outcome. Election night forecasts, which endeavour to forecast the ultimate result before the final outcome is known using early results, were also carried out, with some more accurate than others. After successfully predicting most of the metropolitan region results correctly in the South African local 2016 municipal elections, using an election night forecasting model developed for South Africa (SA), the question of adapting the model to work outside of SA on a different electoral system was raised. The focus of this paper is to describe the results obtained for the 2016 USA presidential election, on election night, using an adapted version of the SA model. This paper also addresses the applicability of the model assumptions as well as the data issues involved in forecasting outside of South Africa. It is shown that even with many hurdles experienced in the process the model performed relatively well.
Getting to grips with election night forecasting: Predicting the unpredictable world of politics
(CSIR, 2009-04-23) Holloway, Jennifer P; Ittmann, H; Greben, J; Elphinstone, C; Schmitz, P; de Villiers, R; Zaid, K
As soon as the election results start coming in, a team of seven CSIR researchers will get to work to do election night forecasting. These scientists will use statistical modelling based on the first actual voting results that are received to do the forecasting. "Our scientific model works - we are confident that we'd be able to once again give an accurate forecast," comments Hans Ittmann, Executive Director of CSIR Built Environment
Integrated land use and transportation modelling and planning: A South African journey
(2020-10) Waldeck, Louis; Van Heerden, Quintin; Holloway, Jennifer P
Confronted by poverty, income disparities and mounting demands for basic services such as clean water, sanitation and health care, urban planners in developing countries like South Africa, face daunting challenges. This paper explores the role of Integrated land use and transportation modelling in metropolitan planning processes aimed at improving the spatial efficiency of urban form and ensuring that public sector investments in social and economic infrastructure contribute to economic growth and the reduction of persistent poverty and inequality. The value of such models is not in accurately predicting the future but in providing participants in the (often adversarial) planning process with a better understanding of cause and effect between different components of the urban system and in discovering common ground that could lead to compromise. This paper describes how an Urban Simulation Model was developed by adapting one of the leading microsimulation models (UrbanSim) originating from the developed world to South African conditions and how the requirements for microscopic data about the base year of a simulation were satisfied in a sparse data environment by introducing various typologies. A sample of results from three case studies in the cities of Tshwane, Ekurhuleni and Nelson Mandela Bay between 2013 and 2017 are then presented to illustrate how modelling supports the planning process by adding elements of rational analysis and hypothesis testing to the evaluation of proposed policies.
Integrating land-cover data with data on population and household characteristics to assess densification along the BRT route in the City of Tshwane
(2018-07) Dudeni-Tlhone, Nontembeko; Petzer, Engela; Gxumisa, Athi A; Holloway, Jennifer P; Makhanya, Sibusisiwe A
This paper is centred on an investigation of whether the integrating earth observation and census data can result in useful information to support transport planning and monitoring. In particular circular buffers with radii of 500 m were created covering the current City of Tshwane’s Bus Rapid Transit (BRT) service areas in order to estimate the proportion of population with convenient access to BRT; and to assess densities in terms of both population and relevant land use characteristics. A combination of data sets were used, namely, the 2011 census, GPS locations of the BRT stops and BRT routes, the satellite derived urban land-cover and the building land use data. The results indicate that 5% of the population in the City of Tshwane has convenient access to the BRT service. Population densities along the BRT service area range from 2 351 to about 37 518 people per 0.79 km2. Some of the BRT service buffers have low population densities and low building densities; while others have moderate to high population densities and high proportions of residential and commercial properties. Routes from Pretoria central towards Sunnyside vary from medium to high densities with respect to population and residential (cluster type residences, flats and student accommodation) and commercial buildings. It can be deduced from these results that there is potential to proceed with proposed densification strategies along BRT routes given that current densities are lower than desired at specific parts of the route. This could subsequently promote public transport accessibility and usage. In conclusion, publically available data that were used in this study and further enhancements of the methodology can be used as a tool for monitoring the implementation of the densification strategy along the BRT route.
Model for election night forecasting applied to the 2004 South African elections
(ORSSA - Operations Research Society of South Africa, 2006-06) Greben, JM; Elphinstone, E; Holloway, Jennifer P
A novel model has been developed to predict elections on the basis of early results. The electorate is clustered according to their behaviour in previous elections. Early results in the new elections can then be translated into voter behaviour per cluster and extrapolated over the whole electorate. This procedure is of particular value in the South African elections which tend to be highly biased, as early results do not give a proper representation of the overall electorate. In this paper the authors explain the methodology used to obtain the predictions. In particular, they look at the different clustering techniques that can be used, such as k-means, fuzzy clustering and k-means in combination with discriminate analysis. The authors assess the power of the different approaches by comparing their convergence towards the final results.
Modelling representative population mobility for COVID-19 spatial transmission in South Africa
(2021-10) Potgieter, A; Fabris-Rotelli, IN; Kimmie, N; Dudeni-Tlhone, Nontembeko; Holloway, Jennifer P; Janse van Rensburg, C; Thiede, R; Debba, Pravesh; Manjoo-Docrat, R; Abdelatif, N; Makhanya, S
The COVID-19 pandemic starting in the first half of 2020 has changed the lives of everyone across the world. Reduced mobility was essential due to it being the largest impact possible against the spread of the little understood SARS-CoV-2 virus. To understand the spread, a comprehension of human mobility patterns is needed. The use of mobility data in modelling is thus essential to capture the intrinsic spread through the population. It is necessary to determine to what extent mobility data convey the same message of mobility within a region. This paper compares different mobility data sources by constructing spatial weight matrices and further compares the results through hierarchical clustering. This provides insight for the user into which data provides what type of information and in what situations a particular source is most useful.
Prediction of the 2004 national elections in South Africa
(Acad Science South Africa, 2005-03) Greben, JM; Elphinstone, C; Holloway, Jennifer P; De Villiers, R; Ittmann, HW; Schmitz, P
During the last three elections in South Africa, the CSIR was involved in the prediction of the final outcome on the basis of early results. In this paper, we describe the methods used by the CSIR in these elections and comment on the success of the model used. We compare the rate of convergence of our predictions towards the final results with the convergence of the actual results. We also comment on the special challenges and time pressures faced by a research team when it uses a scientific analysis tool in a real-time context. The performance of the system is determined by its ability to deliver accurate predictions at an early stage, since interest in predictions diminishes rapidly as the outcome becomes clear. In the event, our predictions proved to be very accurate and our forecasts played a vital role in the ability of the national broadcaster to 'call' the election in the hours after the voting stations closed.
Spatial age-stratified epidemiological model with applications to South African COVID-19 pandemic
(2025) Manjoo-Docrat, R; Abdelatif, N; Holloway, Jennifer P; Dudeni-Tlhone, Nontembeko; Dresselhaus, C; Mbayise, E; Janse van Rensburg, C; Fabris-Rotelli, I; Debba, Pravesh; Makhanya, S
Objectives: This study aims to address the heterogeneity among different age groups in terms of their susceptibility to and transmissibility of infectious diseases. It also seeks to understand how spatial disparities affect disease spread and local population responses to emerging and re- emerging infectious diseases (EIDs and REIDs). Design/Methods: We developed a spatial age-stratified SEIR model using COVID-19 hospitalisation data from South Africa, focusing on the first two waves of the pandemic. This model incorporates contact matrices and demographic data to capture age-dependent and spatial variations in disease dynamics. Results: The spatial age-stratified model produced more biologically plausible and accurate predictions compared to non-stratified models investigated. It highlighted significant differences in COVID-19 risk and transmission across different age groups and regions, offering insights into targeted intervention strategies. Conclusions: The proposed model demonstrates the importance of considering both age and spatial heterogeneity in mathematical models for infectious disease prediction. It provides a valuable tool for governments and public health officials, particularly in resource-limited settings, to develop more effective and targeted interventions. This model can be adapted for other EIDs and REIDs with similar dynamics, enhancing preparedness and response strategies.
A spatial model with vaccinations for COVID-19 in South Africa
(2023-12) Dresselhaus, C; Fabris-Rotelli, I; Manjoo-Docrat, R; Brettenny, W; Holloway, Jennifer P; Abdelatif, N; Thiede, R; Debba, Pravesh; Dudeni-Tlhone, Nontembeko
Since the emergence of the novel COVID-19 virus pandemic in December 2019, numerous mathematical models were published to assess the transmission dynamics of the disease, predict its future course, and evaluate the impact of different control measures. The simplest models make the basic assumptions that individuals are perfectly and evenly mixed and have the same social structures. Such assumptions become problematic for large developing countries that aggregate heterogeneous COVID-19 outbreaks in local areas. Thus, this paper proposes a spatial SEIRDV model that includes spatial vaccination coverage, spatial vulnerability, and level of mobility, to take into account the spatial–temporal clustering pattern of COVID-19 cases. The conclusion of this study is that immunity, government interventions, infectiousness and virulence are the main drivers of the spread of COVID-19. These factors should be taken into consideration when scientists, public policy makers and other stakeholders in the health community analyse, create and project future disease prevention scenarios. Such a model has a place for disease outbreaks that may occur in future, allowing for the inclusion of vaccination rates in a spatial manner.

Browsing by Author "Holloway, Jennifer P"

Results Per Page

Sort Options