MSc - Data Science

University of Hertfordshire

This module is laboratory based, providing students with hands−on training in several techniques fundamental to data science. It provides an environment where students have an opportunity to improve general coding skills, including good practice in software development. All students will create and maintain a git repository for their code and learn the principles of proper version control of software. Through a series of self−contained data science 'challenges' students will learn and put into practice (1) the key principles of data engineering, such as dealing with noisy and missing data, (2) design and implement 10 algorithms and data analysis pipelines to tackle a variety of problems in the most efficient manner, and (3) critically analyse outputs and communicate findings to others.

This module is laboratory based, with an emphasis on the use of three key practical tools for data preparation and processing: Apache Spark, Tensorflow and Keras. Students will gain experience in developing data processing pipelines that exploit these tools and explore how distributed computing and GPUs can be used to scale and accelerate data engineering and processing tasks. Skills will be developed through a combination of hands−on workshops, where fundamental concepts will be introduced, and a series of challenges where students will be expected to develop their own solutions to a range of problems that incorporate the use of Spark, Tensorflow and Keras. The objective is to provide a solid grounding in the use of recognised programming platforms for data engineering and analysis and further develop general programming skills.

In this intensive, non−assessed 'bootcamp' module students will become familiar with the core mathematical, statistical and programming concepts that underpin MSc Data Science. Very little prior knowledge is assumed for this module, which aims to provide a solid grounding in the key basic skills. The curriculum includes: basic algebra and mathematical concepts (e.g. functions), vectors and matrices, linear algebra, differentiation, statistical distributions, probability, and basic programming using the Python language. The delivery of the module content comprises of a set of online lectures, tutorials and problem sets, where students will gain familiarity with the central topics. There are 3 handbooks, one for each topic (maths, statistics and programming), containing notes to guide the learning. Students can check their progress on the module through a series of online formative quizzes. The module will be completed by students in self−directed study.

In this module, which combines lectures with practical workshops, students will learn the key principles and practical skills involved in (a) ingesting, manipulating and processing various forms of data, including binary and ascii formats, and (b) the use of graphic visualisation to explore data sets, for example to discover trends and correlations and other relationships within multi− dimensional data. Students will gain hands−on experience in producing effective visualisation using a range of presentation techniques to communicate findings effectively. This will include the use of versatile practical tools including NumPy, pandas, matplotlib. Students will consolidate learning through practical workshops where real−world data will be used as a basis for applying their skills.

In this module the student will study a selection of topics centred around basic mathematical concepts and skills, typically including basic linear algebra, calculus, basic statistics, fundamentals of probability, basic set theory and information theory. Students will use computational techniques to code some of the mathematical foundation. Students will learn how to select an optimal approach and implement it as a computational algorithm for solving typical data science problems. Actual topics taught may vary from year to year.

You will conduct a software development project as a team. The project will tackle a particular topic in data science that will draw on knowledge and skills developed in the wider programme. You will work in teams to develop team-working skills, collaboration and group study. The module will consist of group and individual assessments such as group reports, group presentations, team discussions and individual tests. There may be some online lectures to provide necessary information such as working as a team, ethics and reviewing literature.

This module will study of a selection of research topics centred around neural network theory and design, machine learning including supervised and unsupervised learning, and some interesting applications, for example, data mining, biocomputation, evolutionary algorithms, neural networks as models of brain function and data visualization. Actual topics taught may vary from year to year.

In this module students will learn how to retrieve and 'mine' information from databases and other large data sources. Students will gain an understanding of SQL and noSQL databases and how to access their contents. Part of the module will focus on data engineering methods such as cleaning that are essential for preparing data for onward processing. Students will also learn how to prepare high−level quantitative and graphic summaries of large datasets. There will be an emphasis on methods to identify trends and anomalies within large datasets, and the use of association rules to discover relationships between data items. Lectures will be supported by practical workshops, where real−world data will be used to enable students to put into practice the techniques discussed in class.

In this module you will explore current data analysis techniques and methods using AI, through the lens of case studies from a range of disciplines such as astronomy, medicine, and finance. You will develop knowledge and skills from studying state-of-the-art research in providing in-depth, downstream analysis, and novel insights of real-world data. It will build on the concepts introduced in the wider programme, including Computer Vision, Natural Language Processing (NLP), and time-series analysis.