The workshop is organised by the Creative Industry Knowledge Transfer Network and was held at the Nesta in London. The workshop had some excellent speakers and I am quite honoured to be invited. Below are the presentation slides.
The paper is accepted by the VAST 2013 conference and will appear in the IEEE Transactions on Visualization and Computer Graphics. It provides visualisation for different uncertainties in Human Terrain Analyisis and a way to construct narratives of the visual exploration discoveries. PDF
With the support from the EU Erasmus funding, I spend a week at the Visual Analytics group led by Prof. Daniel Keim, who is the leading expert on this topic. Konstanz is a nice city in southern Germany, and very busy in summer because of its location and lake. I will definitely want to go back and spend a bit more time there. Below is the talk I gave there about the Data- Intensive Visual Analytics (DIVA) project.
Besides the Middlesex University, there are three other partners in the research consortium: City University (principle investigator), Loughborough University, and Defence Science and Technology Laboratory (dstl).
The project aims to design and develop novel Visual Analytics techniques to address the data quality and uncertainty issues commonly encountered in Social Media analysis. Examples of such issues include the reliability of data source, errors and missing data, artefacts introduced by the analysis algorithms, and human bias in the decision making process. An important part to the solution will be the ‘provenance’, which is the detailed information of the entire decision making life cycle from data collection to reasoning and sense making. Provenance is expected to play a key role in identifying and tracking the data quality and uncertainty issues mentioned previously.
The initial application will be Human Terrain Analysis that provides Army troops with social insights into local population by integrative analysis of social media (e.g., Twitter) together with a range of other intelligence information (e.g., demographics). The results of Human Terrain Analysis will allow the ground troops to choose the appropriate engagement strategy with local communities and avoid any unnecessary confrontation and loss of life. The research outcomes are expected to later extend to other domains such cyber security, finance, and public health.
2011-, adviser, “Spatial-Temporal Analysis of Online Social Information and its Provenance using Visual Analytics”, School of Engineering and Information Sciences, Middlesex University, UK.
Program Committee member, Fifth International Workshop on Data Quality in Integration Systems, in conjunction with the 17th International Conference on Database Systems for Advanced Applications (DASFAA 2012), April 2012, Busan, South Korea.
B.L. William Wong, Kai Xu, and Andreas Holzinger, “Interactive Visualization for Information Analysis in Medical Diagnosis”, Workshop on Human-Computer Interaction & Knowledge Discovery and Data Mining (HCI-KDD) @ 7th Conference of the Austrian Computer Society Workgroup: Human-Computer Interaction (USAB), Graz Austria, November 2011. PDF
Qing Liu, Ji Zhang, James Forbes, Kai Xu, Dinesh Nair, “A Knowledge Discovery Service System for Provenance Exploration”, International Conference on Data and Knowledge Engineering (ICDKE) 2011, Milan, Italy, September 2011. PDF
In this paper we discuss the various aspects in intelligence analysis relating to provenance and the new requirements resulting from the changed nature of terrorist activities (extremely versatile). We propose a three-layer provenance model and a provenance reasoning workspace framework for the design of visual analytics tools for intelligence provenance. PDF
Program Committee member, Fourth International Workshop on Data Quality in Integration Systems, in conjunction with the 16th International Conference on Database Systems for Advanced Applications (DASFAA 2011), April 2011, Hong Kong, China,