Background of Data visualization

说明：这是我写的一篇背景数据可视化背景知识介绍文章，非常简单，为以后论文写作准备的素材。

 

Project topic: Data visualization

1.1 the definition of data visualization

In the subjects of information system and information management, the most basic model is DIKW hierarchical model (Data, Information, Knowledge, Wisdom, DIKW[1]). In data science, this model shows the processing that data is enhanced into information and then into knowledge. What is more, with the help of the quick visual awareness of human eyes and the intelligence cognitive ability of human brain, visualization could play an irreplaceable role in the processing of DIKW model. Modern data visualization use various techinques include computer graphic, image processing, interactive techniques and so on. It could show the valuable information to users by converting acquisition or simulative data into identifiable graphic symbol, chart, diagram and animation.

Data visualization is an active and critical aspect for many domains of subject, such as computer graphic, interactive techniques, data govemance, data analysis,data mining, database and data warehouse, infographic design, and other specific fields.

1.2 Classification and Processing of data Visualization

Scientific visualization, information visualization and visual analytics are the three main branches of data visualization[2]. Firstly, scientific visualization is oriented to science data. And scientific visualization is the oldest and most sophisticated areas in visualization. Secondly, information visualization is oriented to abstract and unstructured data collection. It is wildy used in nearly all areas of human scientific research. Thirdly, Visual analytics is a new-rising subject, which is developed with data mining, data analysic and other data sciences. Essentially, visual analytic is visually bidirectional conversion between machine intelligence and human intelligence, the whole exploration process is iterative, rising in a spiral process. Afterall, there is no clear boundary between these three main branches. Scientific visualization, information visualization and visual analytics collaborate with each other, and the result we called visual practice.The three Figures below show the process of data visualization in different definitions.

 

 

 

1.3 the challenges for data visualization

The definition of visualization is evolve with technology in history. In the modern sence, visualization is a technology when computer and computer display methods develop to a certain stage. When projecter designs visual system, he must content three basic constraints: computing power, human cognitive ability and visual display terminal capability. Therefore, the designers of data visualization face on the challenges include but not limited to:

• computing capacity scalability
• The limitations of perception and cognition
• The limitations of visual display terminal capability
• People-centered exploratory visual analytics [6]
•  Big data visualization and visual analytic[7]

1.4 Current situation of my research area

My research pay most attention on visual analytics. Scientific development and engineering practice in the new period of history shows that the difference between knowledge acquired by intelligent data analysis and knowledge acquired by human is the golden mine of new knowledge. The expression, analysis, and inspection of these differences must make full use of human intelligence, and visual analytics is almost the only feasible way. However, the basic theory and method of visual analytics are still not fully found, it is worth to make further research.

An IEEE Conference on Visual Analytics Science and Technology (VAST) 2012 [6] collected related research topics of visual analytics. The topics include but not limited to:

• Visual expression and interaction techniques, include statistical graphics, information display principle, new pattern of visual presentation, user interface;
•  Data management and knowledege expression;
• Analytical reasoning, include expression of knowledge, knowledge discovery method, perception and cognition;
• Application of visual analytics;
• Evaluation methods;
• Security and privacy;
• Theoretical basis of data transformation in interactive visual analytics;
• The basic algorithm and technology of data visualization, include adaptability of the equipment, web interface, mobile equipments and so on.

 

In this area, data mining and data visualization go hand in hand.  Finding complex patterns in data and making them visible for further interpretation utilizes the power of computers, along with the power of the human mind. Used properly, this is a great combination, enabling efficient and sophisticated data crunching and pattern recognition. So some data sicenctists propound visual data mining[8], its core is presenting the result of data mining with visualization methods. But visual data mining is still using machine intelligence to mine valuable data, not in visual thinking.

Visual analytics must base on data. Our analysis objects are often hypermedia data, such as image, video, sound, hyperlink, rich text, society network, location information. Just refer to the case of mricoblog, for the photographs we could make images grid to show the similar collection of people; For the video we could make video visual summary to show the characteristics pattern with less information; For the text we could make word cloud to show the text analysis result; For the society network, we could make treemap or network chart to explore users’ relationship. In conclusion, different data types have different visualization methods and technology, we simply summed up as follows[10]:

1. Spatial data visualization
2. Geographic information visualization
3. Time-varying data visualization
4. Hierarchical data visualization
5. Network data visualization
6. Text data visualization
7. Media data visualization
8. High dimensional multivariate data visualization

 

Considering the usable range of visual analytics is too large, we need to focus on the specific domain that we interest. The most interesting areas that I concert in visual analytics are two part: one is big data visual analytics, especially video information sampling and analysis; the other is visualization and data analysis of personal digital life.

However, no matter what kind of domain, we will face with the same visual problem: how to make a tool which could automatic filter and aggregate data with a visual interface for users searching for interesting patterns or anomalies?  The big problem can be broken up into four small problems:

• How to automatically filter and aggregate data with user’s order or not?
• How to make an data visualization application with existing technology?
• How to design customer satisfactory interaction in the visual application?
• How to measure the success of a visual practice?

References:

[1]Jennifer Rowley, The wisdom hierarchy: representations of the DIKW hierarchy. Journal of Information Science, 33(2),2007:163-180

[2]Matthew Ward, Georges Grinstein, Daniel Keim, Interactive Data Visualization: Foundations, Techniques, and Applications. May, 2010.

[3]R.B Haber, D. A. NcNabb. Visualization idioms: A conceptual model for scientific visualization systems. Visualization in Scientific Computing,1990:74-93

[4]Ross Bunker, Jock Mackinlay, Robert Morton, and Chris Stolte: Dynamic Workload Driven Data Intergration in Tableau[2009].

[5]  D Keim, G Andrienko, JD Fekete, C Görg, Visual analytics: Definition, process, and challenges – 2008

[6] Visual Analytics Science and Technology (VAST), 2012 IEEE Conference on,

http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=6381601

[7] McKinsey Global Institute, J. Manyika, and M. Chui. Big Data: the next frontier for innovation, competition and productivity. 2011

[8] Michel Crampes, Jeremy de Oliveira-Kumar, Sylvie Ranwez, Jean VillerdVisualizing Social Photos on a Hasse Diagram for Eliciting Relations and Indexing New Photos 2009

[9] C Yu, D Yurovsky, TL Xu - Infancy, Visual data mining: An exploratory approach to analyzing temporal patterns of eye movements, 2012 - Wiley Online Library

[10] 陈为等，数据可视化，电子工业出版社 2013