This three-year, part-time programme is ideal for graduates and working professional engineers who want to advance their careers through data analytics, specifically in the processing industry. Gain the skills to leverage process data for actionable insights by developing a fundamental understanding of dynamic processes and the machine learning methods used to analyse them.

The hybrid online delivery mode is ideal for part-time students worldwide. Whether you work in chemicals, minerals, paper and pulp, food and beverage, pharmaceuticals, or any other processing industry, this programme is designed to help you build a learning community that supports you in applying your newly developed skills. The language of instruction is English.

This programme is designed to equip engineers with the fundamentals of data science, specifically machine learning, enabling them to apply new methods in large, integrated industrial processes. This course focuses on the process industry, covering topics such as plantwide dynamics and advanced process monitoring and control. By emphasising the fundamentals of machine learning, we encourage engineers to understand the methods they use rather than relying on black-box solutions.

Postgraduate students register for 180 credits over the duration of the programme, including eight taught modules and one research project module. Each taught module consists of 15 credits, amounting to 150 notional hours of work. All modules are presented in a hybrid, synchronous mode, which may be attended in person or remotely, but active real-time participation is required. A brief overview of individual modules is provided in the module overview section. The 60-credit research project (i.e., 600 notional hours of work) is the final step towards earning the degree. The project presents an opportunity to work on a complex, relevant problem in the field of process data analytics, supported by individualised supervision from experts in the field. The hybrid programme emphasises synchronous interactions while supporting fully remote learning and is completed over a period of three years to enable part-time studies. Part-time students will be required to spend approximately 600 hours per year on the programme, including at least 15 full days per year for active participation in interactive lectures and tutorials.

A typical module consists of two weeks of pre-reading, followed by a block week that can be attended online or on campus. The final part of each module is a six-week post-block period where students work on assignments, both collaboratively and individually. Most assessments are in the form of assignments and do not require students to be on campus.

Academic activities during pre-reading block weeks vary between modules, and students must consult the relevant module framework beforehand. These activities typically include reading, online quizzes, or mini-assessments. These sessions are asynchronous, allowing students to engage with the content at their own pace. The same approach applies to post-block weeks, which generally include more comprehensive assessments.

The lecture block week offers students the opportunity to engage critically with lecturers and peers on fundamental concepts and applied problems. Attendance during the block week is not required in person at the Stellenbosch campus; it is offered in a hybrid mode, accommodating both in-person and remote students. However, the block week is a synchronous session, requiring active participation in real-time, whether on campus or online. Students are expected to attend the full block week during normal working hours, and part-time students will likely need to apply for study leave to do so.

This NQF level 9 programme consists of set modules, with a minimum number of credits to be completed each year. Students register for a total of 180 credits over three years, including eight taught modules and one research project module.

Each taught module is worth 15 credits, with one credit representing 10 notional learning hours. “Notional learning hours” is the estimated time required for the average student to achieve the specified outcomes of the module or programme. Each taught module therefore amounts to 150 notional hours of work. The research project in the third year is a 60-credit module, equating to 600 notional hours.

Data Science (Eng) (14190-874) (15 credits) (Year 1, term 1)

Why this module matters

Data science involves applying computational, statistical, and machine learning techniques to gain insights into real-world problems. This module focuses on the data science project life cycle, providing a clear understanding of the five steps in the data science process: obtain, scrub/wrangle, explore, model, and interpret. These steps form the foundation for all data science investigations.

Module outcomes

You will gain an appreciation of the requirements, complexities, and tools needed for each step of the project life cycle. You will also understand the process of constructing a data pipeline, from raw data to knowledge. Case studies from the engineering domain will be used to explore each of these steps.

Applied Machine Learning (14022-874) (15 credits) (Year 1, term 2)

Why this module matters

Open-source software libraries have made a vast array of machine learning tools readily available. However, a fundamental understanding of these techniques is essential for correct implementation and interpretation of results. This module aims to develop that understanding, focusing on information-based learning, similarity-based learning, error-based learning, kernel-based learning, probabilistic learning, ensemble learning, and incremental learning.

Module outcomes

You will gain practical experience in implementing a wide range of machine learning techniques. Not only will you learn the theoretical underpinnings of several machine learning techniques, gaining an important understanding of the requirements, inductive bias, advantages, and disadvantages, but you will also acquire the practical know-how needed to apply these techniques to real-world problems.

Plantwide Control (10894-872) (15 credits) (Year 1, term 3)

Why this module matters

Process engineers must understand the effects of plant dynamics and control to successfully implement any form of intervention. This module provides an overview of plant-wide control and advanced process control, focusing on understanding cause-and-effect relationships in systems with multiple unit operations, the effects of process integration on dynamics, and advanced control strategies specific to integrated systems. This knowledge will support the interpretation of process data analysis results.

Module outcomes

You will be equipped to critically engage with the process control field of specialisation. This includes the evaluation of control strategies, interpretation of common representations of system dynamics, and understanding the advanced process control hierarchy. Importantly, you will develop the ability to critique plant-wide control schemes, accounting for the complex system dynamics resulting from the high interconnectivity of unit operations.

Advanced Topics in Engineering Management (11748-873) (Year 1, term 4)

Why this module matters

The sustainability of any solution or intervention (including data-driven solutions) depends on its integration into existing business processes. The purpose of this module is to present principles of general management within the context of technical disciplines. The course themes include the business environment and strategic management at the firm level, touching on the role of innovation and technology for competitiveness at both the systems level and from international and national perspectives.

Module outcomes

You will focus on tools and techniques for technology and innovation management, exploring the link between technology management and business management from a capabilities approach. The functions of engineering management – namely planning, organising, leading, and controlling – will also be discussed, with a specific focus on human resource management.

Dynamic Data Analytics (10884-872) (15 credits) (Year 2, term 1)

Why this module matters

The analysis of dynamic time-series data presents a variety of challenges but is well supported by classical control theory as well as new developments in machine learning. This module provides a practical understanding of both traditional and contemporary approaches to analysing and leveraging data produced by measurements of dynamic processes specific to the processing industries, including systems identification, data reconciliation, and state estimation.

Module outcomes

You will develop a conceptual understanding of the complexities, limitations, and opportunities associated with time series data. You will gain experience in systems identification, state estimation, and data reconciliation, and explore research at the intersection of machine learning and process control, understanding the impact of recent innovations on traditional techniques, including regularisation and kernel methods. This will equip you to critically evaluate process datasets and identify appropriate analysis methods for specific use cases.

Numerical Methods (36323-876) (15 credits) (Year 2, term 2)

Why this module matters

Numerical methods are central to machine learning, and practitioners need an appropriate understanding of the associated numerical challenges to avoid common pitfalls. These challenges are discussed within the context of matrix computations, which not only illustrate many of the numerical issues routinely encountered but also develop the fundamentals of linear algebra essential for effectively navigating the field of machine learning.

Module outcomes

You will study the effective solution of linear systems, involving both square and rectangular matrices (least-squares). Direct as well as iterative methods are considered, with the emphasis on sparse matrices and matrices with structure. Numerical methods for the eigenvalue problem are also considered. Pitfalls such as numerical instability and ill-conditioning are pointed out. Theory, algorithmic aspects, and applications are emphasised in equal parts.

Optimisation (Eng) (14020-874) (15 credits) (Year 2, term 3)

Why this module matters

Almost all machine learning methods rely on some form of optimisation, and the ability to build and solve an optimisation problem is essential for the effective use of machine learning techniques for data analysis. This module will explore common classes of optimisation problems and the algorithms required to efficiently solve them.

Module outcomes

You will gain experience in identifying, formulating, and solving important classes of optimisation problems, both by hand and using appropriate software, and analyse results using techniques such as sensitivity analysis. More importantly, you will learn to apply a deep conceptual understanding of optimisation methods to solve practical machine learning problems.

Integrated Data Analytics (10884-873) (15 credits) (Year 3, term 1)

Why this module matters

Combining and applying knowledge and methods from various domains to real-world problems remains a significant challenge. Critically evaluating a selection of case studies from industrial processing plants will inform common best practices and provide an opportunity for both lecturers and students to share experiences. The case studies will also serve as a framework to prepare you for the research project required for the chemical engineering programme.

Module outcomes

You will be able to implement the full process data analysis workflow, including exploratory data analysis, identification of integrated process units, application of advanced analytical techniques from multiple specialised disciplines, and synthesis of results. Additionally, you will be equipped to identify appropriate levels of intervention and understand the requirements for integrating data-driven interventions within business processes, specifically within the context of chemical and/or minerals processing plants.

Chemical Engineering Research Project (10882-872) (60 credits) (Year 3)

Why this module matters

You will conduct a major research investigation under the direct supervision of an experienced researcher or industry practitioner in the field of process data analytics. This module serves as the capstone of the postgraduate programme.

Module outcomes

You will demonstrate the ability to identify, formulate, and conduct a research project directly related to the analysis of process data. You will be required to draw on all prior learning to successfully analyse the process using a data-driven approach to generate practical insights.

Students register for 180 credits across the programme, which includes eight taught modules and one research project module. Each 15-credit taught module involves two weeks of pre-reading, a block week (attended online or on campus), and a six-week post-block period for assignments. Most assessments are assignments and do not require campus attendance.

Please refer to the 2025 timetable below.

All applicants must meet the minimum admission requirements as detailed in the admission requirements section (also specified in Section 3.6 of the Engineering Calendar, Part II).

For the Data Analytics focus area, the minimum selection criterion is a BEng or BSc (Eng) degree (NQF level 8) in a discipline that includes prior exposure to control systems, typically chemical, mechanical, or electrical/electronic engineering.

The following documents should be included in your application:

• Complete academic record(s),
• Degree certificate(s),
• Comprehensive curriculum vitae, and
• Motivation letter: Applicants must upload a one-page motivation letter detailing their prior learning and/or industrial experience relevant to the programme’s minimum admission requirements, specifically exposure to control systems. Relevant prior learning may include undergraduate training in chemical, mechanical, or electrical/electronic engineering, or related short courses. If relevant prior learning is lacking, applicants should provide a thorough description of their industrial experience.

Complete and submit an institutional application. After submission, the Central Admissions Office will review the application. If the submitted documents are incorrect, candidates will be notified through the applicant portal, and their application status will be updated to “incomplete”.

After the programme’s application deadline, all applicants who meet the criteria will undergo a selection process reviewed by a departmental committee. This committee evaluates applicants based on their academic records and relevant industry experience. The list of candidates who meet the selection criteria will be presented for approval at a departmental management meeting. All applicants will be notified of the outcome before the end of the application year.

Successful candidates will receive a conditional offer through the SU applicant portal, which they must accept. Following this, a final offer will be issued, which candidates will also need to accept.

Please note that international students pay international student fees, which may vary based on the candidate’s primary citizenship. This page contains useful information about international student fees.

A personalised provisional quote can be requested here.

Should you require further information or support regarding tuition fees, please send an email to [email protected] and we’ll respond as soon as possible.

The Data Analytics focus area involves a significant amount of mathematical modelling, statistics, and coding. Throughout the programme, we will use Python as the primary environment within which we engage with data science and machine learning.

It is often difficult to assess your own level of preparedness. We suggest that you browse through the book Mathematics for Machine Learning by Deisenroth, Faisal, and Ong. The associated GitHub page contains tutorial notebooks on Linear Regression, Principal Component Analysis, and Gaussian Mixture Models. Open the Jupyter Notebooks in Google Colab by clicking the appropriate links and look through them: they contain a good mixture of linear algebra, statistics, and code. We don’t necessarily expect you to understand the presented mathematics and statistics before you start the programme, but this will definitely be the type of work you will engage with, and we will certainly require you to be able to complete these tutorials by the end of the first year. You should also be able to understand most of the code in the Solutions notebooks (see, e.g., the Linear Regression Tutorial Solution notebook), if not necessarily the statistics.

If the mathematics and statistics presented in the book and the notebooks seems interesting and exciting to you, and you have a sense of what the code represents, then this programme is for you.

If you are interested but your concerned you are not suitably prepared, we recommend you take a short refresher course before starting. I personally like the freely available textbook Python Programming and Numerical Methods by Kong, Siauw and Bayen. If you can complete the first twelve chapters of the book, you will be in a very good position to engage with the learning material. Similarly, the famous introductory textbook An Introduction to Statistical Learning by James et al. includes applications in Python. This is a good alternative to start learning Python and the basics of statistical learning simultaneously.

If you would prefer a more formal refresher on coding, consider completing one of the following courses before starting with the Data Science module in 2025:

• MITx: Introduction to Computer Science and Programming Using Python
• CS50’s Introduction to Computer Science
• Expressway to Data Science: Python Programming Specialization

These recommendations are not to be taken lightly:
a working knowledge of Python is required for the first module you will be enrolled for.

For any queries or support, please send an email to [email protected].