Unit information: Visual Analytics in 2020/21

Unit name	Visual Analytics
Unit code	COMSM0038
Credit points	20
Level of study	M/7
Teaching block(s)	Teaching Block 2 (weeks 13 - 24)
Unit director	Professor. Nabney
Open unit status	Not open
Pre-requisites	Software Development: Programming and Algorithms
Co-requisites	EMATM0044
School/department	Department of Computer Science
Faculty	Faculty of Engineering

Description including Unit Aims

Visual analytics couples the visual representation of data with analytical processes to support complex decision making and understanding. A picture may be worth a thousand words, but only if it is well designed to represent data faithfully and meaningfully. This unit will enable students to create powerful analyses of data and communicate them effectively to non-specialists.

This unit will cover two key aspects of visual analytics: the science of information visualisation (primarily concerned with the way that data is represented visually), and advanced machine learning (as a tool to change the data representation, e.g. through dimensionality reduction, or as a way of analysing visual data) in a framework of statistical pattern recognition.

Information visualisation topics covered by this unit include: data types and their representations, non-vectorial data, human requirements for visual analytics, scientific visualisation, visualisation quality metrics, Shneiderman’s mantra (overview first, zoom and filter, details on demand) practical visualisation tools.

Machine learning topics covered by this unit include: principles of Statistical Pattern Recognition (probabilistic models for data, curse of dimensionality generalisation error, bias-variance dilemma); linear models (Probabilistic Principal Component Analysis; Discriminant Analysis); generalised dissimilarity mappings and neighbour embedding techniques; Gaussian Processes; latent variable models (Gaussian Mixture Models, Generative Topographic Mapping and Gaussian Process Latent Variable Model); Bayesian model regularisation and combination; feature selection; challenges of large datasets and potential solutions.

Throughout the unit there is a focus on understanding theory and modelling principles in order to apply them effectively to represent data.

Intended Learning Outcomes

Students will be able to

1. Define and apply the principles of information visualisation.
2. Analyse the design of visual representations of data in terms of human perception and cognition.
3. Define the types and semantics of data.
4. Build machine learning models for data and explain their operation in terms of a statistical pattern recognition framework.
5. Use Bayesian regularisation and variational methods to fit models.
6. Create user-focused visualisations of numerical, categorical, time series, and network data using Python visualization tools and Tableau.

Teaching Information

Teaching will be delivered through a combination of synchronous and asynchronous sessions, including lectures, practical activities and self-directed exercises.

Assessment Information

80% coursework, 20% in-class tests.

Reading and References

• Tamara Munzner. Visualization Analysis and Design, CRC Press, 2014.
• Edward R Tufte. The visual display of quantitative information. Vol. 2. Cheshire, CT: Graphics press, 2001.
• Ware, Colin. Visual thinking: For design. Elsevier, 2010.
• Kyran Dale. Data Visualization with Python and JavaScript: Scrape, Clean, Explore & Transform Your Data, O-Reilly, 2016.
• Christopher M Bishop. Pattern Recognition and Machine Learning, Springer, 2006.
• Ian T Nabney. Netlab: Algorithms for Pattern Recognition, Springer, 2004.