Definition of Data Visualisation
Data visualisation is used in three main phases of the statistical modelling or data analysis process: exploratory data analysis, diagnostics/model checking, and communicating outputs/insights.
In a big data context, difficulties for exploratory data analysis and diagnostics plots arise due to computational limitations or because current visualisation methods are not applicable to datasets that contain different data structures. With a large or complicated data set, generating a simple box plot, a scatter plot or histogram can become computationally costly, time consuming or may not enable an exploration of patterns across data types. ACEMS is currently exploring novel methods for adapting standard visualisation techniques to address these difficulties.
Visualisations intended to communicate statistical insights and outputs represent different challenges depending on the audience, findings and type of data. ACEMS is currently developing methods for communicating statistical uncertainty to non-expert audiences, as this represents an important challenge when dealing with big data problems.
Who works in Data Visualisation
Principal Research Fellow: Dr Tomasz Bednarz
What Expertise do we have?
- High-performance visualisation utilising heterogeneous architectures
- Web based visualisation
- Design approaches to visualisation
- Interactive and immersive visualisation techniques
- Communication of complex data to more general audiences
Our focus areas:
- Measurement Science in Immersive Environments – distribution of variables, spatial information.
Uncertainty – visualisation of uncertainty in data.
- Integration – connecting different visualisation modalities, web vs desktop, unified plaform to drive scientific visualisations.
- Focus on new hardware modalidies for display input and output and communication: Oculus Rift, Tablets, information Augumentation, 3-D projectors.
- Collaboration frameworks in virtual environments.
- Views of the data – multiple ways to visualise the same data, carry our Human-Computer Interacton experiments to find best communication modality depending on target audience.
- Visual Analytics and Big Data Analytics.
- Modern GPU based rendering techniques, imaging and compute.
Case studies: Concrete examples of what we have done
Computational Simulation Sciences:
Links to other capabilities: (e.g., Design / Optimisation)
- diagnostics, diagnostic plots
- information design
- model checking
- exploratory data analysis (EDA)
- initial data analysis (IDA)
- graphs, graphical tools, graph
- data visualisation, data visualization
- immersive visualisation
- visual analytics
Dr Tomasz Bednarz (email@example.com)