Data management, integration, analytics, mining and visualisation with a particular interest in life sciences and health informatics applications. While my own research has focused on techniques appropriate for life sciences applications, I am interested in wider application areas and have supervised projects involving the management and analysis of data in areas such as archaeology, astronomy, genealogy, geography and history. As I enjoy music and play a number of instruments with enthusiasm if not great technique, I am always interested in applications involving music.
In the biological sciences, next generation sequencing technologies give rise to particular problems in identifying and visualising significant features within the data volumes generated. For example, data is generated across multiple time points for multiple patients for a virus that is actively mutating, and analytic techniques are essential to understanding and interpretion of the data in immunology and epidemiology studies. One possibility is to represent relationships between the data as graphs. Can you develop software to support the analysis and visualisation of such graph data?
More generally, with "big data" it becomes increasingly difficult for users to understand the patterns or relationships in the data which represent interesting information. Can you take a complex data set and develop and implement visualisation techniques which enable interesting information to be more easily recognised and understood by users?
Many applications require access to "related" data in different databases but it is not pre-defined what the relationships between the databases are as would be the case in a conventional data integration setting. Techniques are emerging to enable querying of linked databases where only partial information is available on the nature of the links, and this information is changing over time. Can you develop and implement techniques which support the querying, analysis and mining of such linked data sources?
Alternatives to the classical relational database architecture, sometimes referred to as NoSQL, are often characterised by not having a fixed schema nor a record-oriented underlying structure. Traditional database capabilities such as querying, optimization, concurrency control, as well as more advanced capabilities such as data analytics, data mining and database integration, are not supported in the same way as in conventional relational databases. Can you take a NoSQL architecture, a database capability and a suitable application area to enable you to assess different approaches to the implementation of that capability for the given architecture?