Mastering the Information Age - Solving Problems with Visual Analytics

 Mastering the Information Age - Solving Problems with Visual Analytics

Edited by Daniel Keim (Scientific Coordinator of VisMaster), Jörn Kohlhammer (Coordinator of VisMaster), Geoffrey Ellis and Florian Mansmann

http://www.vismaster.eu/wp-content/uploads/2010/11/VisMaster-book-lowres.pdf

每章节的概略：

 

Chapter 2 describes some application areas for visual analytics and puts the size of the problem into context, and elaborates on the definition of visual analytics. The interdisciplinary nature of this area is demonstrated by considering the scientific fields that are an integral part of visual analytics.

 

Chapter 3 reviews the field of data management with respect to visual analytics and reviews current database technology. It then summarises the problems that can arise when dealing with large, complex and heterogeneous datasets or data streams. A scenario is given, which illustrates tight integration of data management and visual analytics. The state of the art section also considers techniques for the integration of data and issues relating to data reduction, including visual data reduction techniques and the related topic of visual quality metrics. The challenges section identifies important issues, such as dealing with uncertainties in the data and the integrity of the results, the management of semantics (i.e., data which adds meaning to the data values), the emerging area of data streaming, interactive visualisation of large databases and database issues concerning distributed and collaborative visual analytics.

 

Chapter 4 considers data mining, which is seen as fundamental to the automated analysis components of visual analytics. Since today’s datasets are often extremely large and complex, the combination of human and automatic analysis is key to solving many information gathering tasks. Some case studies are presented which illustrate the use of knowledge discovery and data mining(KDD) in bioinformatics and climate change. The authors then pose the question of whether industry is ready for visual analytics, citing examples of the pharmaceutical, software and marketing industries. The state of the art section gives a comprehensive review of data mining/analysis tools such as statistical and mathematical tools, visual data mining tools, Web tools and packages. Some current data mining/visual analytics approaches are then described with examples from the bioinformatics and graph visualisation fields. Technical challenges specific to data mining are described such as achieving data cleaning, integration, data fusion etc. in real-time and providing the necessary infrastructure to support data mining. The challenge of integrating the human into the data process to go towards a visual analytics approach is discussed together with issues regarding its evaluation. Several opportunities are then identified, such as the need for generic tools and methods, visualisation of models and collaboration between the KDD and visualisation communities.

 

Chapter 5 describes the requirements of visual analytics for spatio-temporal applications. Space (as in for example maps) and time (values change over time) are essential components of many data analysis problems; hence there is a strong need for visual analytics tools specifically designed to deal with the particular
characteristics of these dimensions. Using a sizeable fictitious scenario, the authors guide the reader towards the specifics of time and space, illustrating the involvement of various people and agencies, and the many dependencies and problems associated with scale and uncertainties in the data. The current state of the art is described with a review of maps, geographic information systems, the representation of time, interactive and collaborative issues, and the implication of dealing with massive datasets. Challenges are then identified, such as dealing with diverse data at multiple scales, and supporting a varied set of users, including non-experts.

 

Chapter 6 highlights the fact that currently most visual analytics application are custom-built stand-alone applications, using for instance, in-memory data storage rather than database management systems. In addition, many other common components of visual analytics applications can be identified and potentially built into a unifying framework to support a range of applications. The author of this chapter reviews architectural models of visualisation, data management, analysis, dissemination and communication components and outlines the inherent challenges. Opportunities and next steps for current research are subsequently identified which encourage a collaborative multidisciplinary e ort to provide a much needed flexible infrastructure.

 

Chapter 7 discusses visual perception and cognitive issues - human aspects of visual analytics. Following a review of the psychology of perception and cognition, distributed cognition, problem solving, particular interaction issues, the authors suggest that we can learn much from early application examples. Challenges identified, include the provision of appropriate design methodologies and design guidelines, suitable for the expert analyst as well as the naive users; understanding the analysis process, giving the user confidence in the results, dealing with a wide range of devices and how to evaluate new designs.

 

Chapter 8 explains the basic concept of evaluation for visual analytics, highlighting the complexities for evaluating systems that involve the close coupling of the user and semi-automatic analytical processes through a highly interactive interface. The exploratory tasks associated with visual analytics are often open ended and hence it is dicult to assess the e ectiveness and eciency of a particular method, let alone make comparisons between methods. The state of the art section outlines empirical evaluation methodologies, shows some examples of evaluation and describes the development of contests in different sub-communities to evaluate visual analytics approaches on common datasets. The authors then argue that a solid evaluation infrastructure for visual analytics is required and put forward some recommendations on how to achieved this.

 

Chapter 9 summarises the challenges of visual analytics applications as identified by the chapter authors and presents concrete recommendations for funding
agencies, the visual analytics community, the broader research community and potential users of visual analytics technology in order to ensure the rapid advancement of the science of visual analytics.