Postgraduate Diploma in Data Science

Purpose

The purpose of the Postgraduate Diploma in Data Science is to enable working professionals to undertake advanced study in the field of data science. Data-driven decisions are at the centre of a range of business functions and as such, the field of data science is at the intersection of business domains, coding, mathematics, and statistics. The qualification has a strong interdisciplinary character and is relevant to graduates in applied mathematics, finance, computer science, economics, and other related disciplines.

The qualification will equip learners with the necessary skills to extract, clean, prepare and pre-process data for feeding into the best-suited algorithm(s), and finally for presenting the data in appropriate visual and written formats. The qualification will provide learners with theoretical knowledge and practical experience of the widely used supervised and unsupervised learning algorithms. Learners will be taught to present the final information in the format of a fit-for-purpose report, by depicting results in appropriate tables, supported by an interpretation of results. These skills, combined with domain knowledge, will provide the prospective data scientist with the necessary skills to provide practical solutions and future predictions to complex numerical problems in business, and any field where data is used extensively.

On completion of the qualification, successful learners will be able to

• Apply different data exploration methods to gain a critical understanding of data, variables, and observations.
• Draft a data science project plan to apply decision-making skills to solve complex problems.
• Demonstrate understanding of data science methods and techniques.
• Apply data management skills to code, prepare, pre-process, manipulate, and visualise data in preparation for the model-building phase.
• Build, fit, and tune models with respect to different data science methods.
• Interpret and critically evaluate solutions to which a particular data science method has been applied in practice.
• Apply the built models on unused data and use domain knowledge to provide decision-makers with accurate predictive information to enable the formulation.

Rationale

South Africa is inundated with data, constantly moving, and renewing itself in a data-driven world. Society needs personnel who can apply critical, analytical tools and algorithms to model and explain patterns observed in data using mathematical and statistical techniques. This qualification will enable the learner to work with data using Microsoft Excel and Python, manage and use data, and apply a variety of statistical learning algorithms such as Logistic Regression, Support Vector Machines, Decision Trees, Boosting Algorithms, Neural Networks, K-Means Clustering and Hierarchical Clustering. Learners will furthermore gain knowledge in the domain area of either finance or marketing via the elective module, to enable them to function optimally as data scientists in these fields.

The qualification will enable the individual working in most sectors or departments such as Financial, Banking, Insurance, Marketing, Tourism, Meteorological, Health, Information Technology (IT) and Business, to drive data-driven decisions for those holding managerial positions. There are many Data Science jobs advertised on a multitude of platforms. The enormous amount of data available needs to be mined, cleaned, fitted, trained, and evaluated using appropriate algorithms so that key decisions and future predictions can be made. From banking to transport and shopping, everyday activities are increasingly leaving digital footprints that are transforming society's households and workplaces. The richness of data is changing the dynamics of many professions, and employers are increasingly seeking workers who can help them make sense of it.

The qualification is ideal for learners with a strong proficiency in analytical skills who have an affinity and feel for making sense of large amounts of data. The candidates will gain experience in reading, cleaning, exploring, fitting, training, and evaluating large structured and unstructured data sets to reduce the data into a usable form so that key decisions can be extracted from the raw information. Successful candidates will be able to operate in any field or sector where large amounts of data are available, which includes most areas of industry such as banking, insurance, finance marketing, tourism, meteorological, health, IT, business, municipalities, and housing.

The qualification will enable learners to pursue the following careers

• Data Scientist.
• Data analyst.
• Data engineer.
• Statistical Learning Specialist.
• Machine Learning Specialist.
• Senior Data Scientist.
• Specialist Data Scientist.
• Principal Data Scientist.
• Data and Analytics Manage.
• Python Data Scientist.

Data Science is one of the most popular emerging domains and most sought-after career options. Data Science is revolutionising almost every industry and dominates emerging job rankings. The fourth industrial revolution (4IR) with an enormous amount of data available and exponential increases in computing power appropriately qualified individuals need to mine this data to identify hidden patterns and extract actionable insights. Managers and decision-makers in most sectors rely on qualified personnel with well-developed analytical skills to analyse, present and communicate results to optimise and improve productivity.

The most appropriate learning pathway would reside in any one of finance, marketing, business, or information technology. A background in statistical, mathematical or computer science would be optimal but not a necessity. There are many data scientists who learn the skill of data science through postgraduate degrees, diplomas, or courses. A person with a strong domain knowledge of the sector in which he/she works such as the banking industry forms a prominent pillar in the art of being a data scientist, computer science, statistics, and mathematics.

Society will benefit from these highly sought-after professionals who can effectively operate in the 4IR with the availability of mega data and a continued exponential increase in computing power. The professionals will contribute to society in many ways, from simple and yet important tasks such as traffic control, customer service, online buying habits, advertising, hospital waiting time, predicting the weather and exchange rate to more advanced tasks such as self-driving cars, fraud detection, energy exploration, robotics, and genetic manipulation. Subsequently, the economy will grow in many ways due to improved efficiency, effectiveness, decisions, productivity, and profitability in all sectors.

Recognition of Prior Learning (RPL)

The institution has an approved Recognition of Prior Learning (RPL) policy applicable to equivalent qualifications for admission into the qualification. RPL will be applied to accommodate applicants who qualify. RPL thus provides alternative access and admission to qualifications, as well as advancement within qualifications.

RPL for access

• Applicants who do not meet the stated admission criteria, but who have relevant work experience/prior learning may apply for admission under the policy on Recognition of Prior Learning (RPL).
• The institution admits a maximum of 10% per cohort via RPL.
• The implementation of RPL is context-specific, in terms of discipline, programme and level.

RPL for exemption of modules

• In specified circumstances qualifying applicants may also engage in the RPL for exemption process, where any form of informal, formal, or non-formal learning will be assessed for relevance towards possible module exemption.

Entry Requirements

The minimum entry requirement for this qualification is

• Advanced Diploma in Information Technology, NQF Level 7.

Or

• Bachelor of Computer Science, NQF Level 7.

Or

• Bachelor of Science in Data Science, NQF Level 7

Or

• Bachelor of Science in Information Technology, NQF Level 7.

Or

• Bachelor of Information Technology, NQF Level 7.

This qualification consists of the following compulsory and elective modules at NQF Level 8 totalling 120 Credits.

Compulsory Modules, Level 8, 100 Credits

• Working with Data, 20 Credits.
• Introduction to Data Science and Statistics, 20 Credits.
• Data Science I, 20 Credits.
• Data Science II, 20 Credits.
• Capstone Project, 20 Credits.

Elective Modules, Level 8, 20 Credits (Select one module from the following options)

• Data Science in Finance, 20 credits.

OR

• Data Science in Marketing, 20 credits.

1. Apply different data exploration methods to gain a critical understanding of data, variables, and observations.
2. Demonstrate the ability to draft a data science project plan to apply decision-making skills to solve complex problems.
3. Demonstrate an understanding of data science methods and techniques.
4. Apply data management skills to code, prepare, pre-process, manipulate, and visualise data in preparation for the model-building phase.
5. Demonstrate the ability to build, fit, and tune models with respect to different data science methods.
6. Demonstrate the ability to interpret and critically evaluate solutions to which a particular data science method has been applied in practice.
7. Apply the built models on unused data and use domain knowledge to provide decision-makers with accurate predictive information to enable the formulation of important decisions.

Associated Assessment Criteria for Exit Level Outcome 1

• Analyse the nature and structure of data.
• Apply descriptive statistical techniques to explore the data.
• Apply data exploring and visualising techniques using Python and associated libraries.

Associated Assessment Criteria for Exit Level Outcome 2

• Discuss the data science process.
• Draft a data science project plan.

Associated Assessment Criteria for Exit Level Outcome 3

• Discuss the concepts related to data science and statistics.
• Analyse a range of algorithms available to solve data-related problems.

Associated Assessment Criteria for Exit Level Outcome 4

• Apply appropriate data manipulation techniques to prepare data for analysis.
• Use appropriate tools to apply statistical and mathematical operations to prepare data.
• Produce graphical representations to assist with the preparation, pre-processing, and cleaning of data.

Associated Assessment Criteria for Exit Level Outcome 5

• Explain backpropagation, gradient descent, learning rate and cost functions.
• Build, fit, and train models related to logistic regression, support vector machines, decision trees, boosting algorithms, neural networks, K-Means clustering and hierarchical clustering.
• Assess model performance and variance/bias trade-off.
• Configure and tune the parameters.
• Apply strategies to manage the complexities of these algorithms.

Associated Assessment Criteria for Exit Level Outcome 6

• Evaluate and interpret models related to logistic regression, support vector machines, decision trees, boosting algorithms, neural networks, K-Means clustering, and hierarchical clustering.
• Evaluate the performance of certain model solutions compared to others for a particular data science problem.
• Synthesise information into a well-written report using written and graphical techniques.

Associated Assessment Criteria for Exit Level Outcome 7

• Apply the models to make predictions using unused data.
• Together with domain-specific knowledge, provide bulleted practical solutions and recommendations to the client.
• Present the algorithm ready to use for future predictions.
• Present all information, analytical procedures, and findings to a panel of academics and industry-specific experts.