Spatiotemporal Knowledge Visualization and Discovery in Dynamic Social Networks
9 likes1,002 views
The document discusses spatiotemporal knowledge visualization in dynamic social networks, highlighting the challenges and motivations for effective visualization tools. It outlines the architecture and concepts of the Dynamic Visualization Tool (DYVT) and emphasizes the importance of user customization and data interoperability. The paper concludes with future work on enhancing visualization techniques and exploring more complex social network data applications.
Spatiotemporal Knowledge Visualization and Discovery in Dynamic Social Networks
- 1. Spatiotemporal Knowledge Visualization and Discovery in Dynamic Social Networks Ralf Klamma, Yiwei Cao , Marc Spaniol, Yan Leng Graz, Austria, 5-7 of September, 2007 Special Track on Knowledge Visualization and Knowledge Discovery (KVD ’07 at i-know) Informatik 5, RWTH Aachen University, Germany
- 2. Agenda Introduction Motivation Scenario Related work Social networks and visualization Spatiotemporal knowledge visualization languages DyVT: Dynamic Social Network Visualization Tool System concepts 3-Tier architecture Conclusions and outlook
- 3. Introduction: Visualization Spatiotemporal knowledge visualization Main tasks of geographers and cartographers (Andrienko, 2006) Social network visualization Source: Touchgraph Facebook Browser, TouchGraph LLC, 2007 Source: “A technique for spotting connections” NYTimes, Feb. 25, 2006.
- 4. Motivation “ Structural Bias” (Milgram, 1967) Lack of means of spatiotemporal knowledge visualization of rich dynamic information Temporal attributes Geospatial attributes Limited interchange formats for social network data Text-based formats (UCINET DL, Pajek .net…) XML-based formats (GraphML, DynetML…) Few options or controls by end users
- 5. Scenario Interoperability User’s interactions Unified toolkits Dynamic distributed social network data How to visualize complex dynamic social network data? Appearance Data Geospatial Data Temporal Data Relational Data How to interoperate social network data?
- 6. Related Work: Features of Social Networks Spatiotemporal Data in Social Networks Geographic distribution Dynamic change by time Common features Social network analysis ( Brandes & Erlebach, 2005 ) Emails, Mailing lists, Forums, Blogs, Wikis, ...
- 7. Social Network Visualization 503 projects at www.visualcomplexity.com 6359 people bookmarks this link (as of Sep.4, 2007) Complexed results Most Visited Projects: Windows vs Linux Server (2006) The Strengths of Nations (2006) Data Visualisation of a social network (2007) Neuronal Network (2006) Mammal Supertree (2007) Hierarchical Edge Bundles (2006) A Networked Designer‘s Critical Path (2004) Visualization of Blogspace (2003) New York Subway Map (1972) Structural Bias?
- 8. Spatiotemporal Knowledge Visualization Temporal visualization Animation layout for node-link graphs (Erten et al. 2005) Readability and mental map preservation Geospatial visualization Spatial data collecting, sharing and analyzing Web map services (Google, 2007) Personalized visualization User preference and customization To visualize: icons, sizes, colors, weight, etc. (Krempel, 2005)
- 9. Spatiotemporal Knowledge Visualization Languages in Social Networks Temporal visualization Animation layout for node-link graphs Readability and mental map preservation Geospatial visualization Spatial data collecting, sharing and analyzing Web map services Personalized visualization User preference and customization To visualize: icons, sizes, colors, weight, etc. DyNetML: dynamic network data as sets of time slices KML: used by Google Earth and Google Maps GraphML: published and well-supported Metadata interoperability Integration
- 10. DyVT: System Concepts Relational data Temporal data Geospatial data Appearance data DyVT XML-based Target Language (DyVTML) Multi-media data types Animation Mashup views Map view SVG GIF JPEG
- 11. System Concepts to 3-Tier Architecture Relational data Temporal data Geospatial data Appearance data DyVT XML-based Target Language (DyVTML) Multi-media data types Animation Mashup views Map view SVG GIF JPEG Database tier Data processing tier Visualization tier
- 12. 3-Tier Architecture Intertier data flow Intra-tier data flow Legend Data Processing Tier Database Tier Visualization Tier
- 13. Database Tier : Social Network Data Spatial data Spatial database Temporal data Social network data Mailing lists BBS/Forums Weblogs Social bookmarks Other multimedia social networks (music, pictures, videos …)
- 14. Data Processing Tier : DyVTML Geospatial data Appearance data ADML Temporal data Temporal data Mailing list data
- 15. Data Processing Tier : Social Network Appearance Data Graphic setting interface XML based ADML format XML Schema for ADML
- 16. Visualization Tier: Layout Algorithms Circle Layout KK Layout Set the radius Arrange nodes Set parameters Get initial positions Calculate energy Rearrange the node until it is below the lowest energy Iterate until get lowest energy
- 17. Spatiotemporal Knowledge Visualization Mailing list visualization based on Google Maps
- 18. Conclusions DyVT DyVTML(ADML) Animation Maps T Temporal data Geospatial data Appearance data S A
- 19. Outlook Interchange Formats Multi-relational social networks More database support Appearance data Various output formats Spatiotemporal knowledge visualization & discovery in dynamical social networks Future Work Visualization and discovery Matrix-based forms Application of other layout algorithms Spatiotemporal data mining Evaluation and openness Further evaluation Integration into other tools (e.g. Paladin, pattern based Dynamic NA) Extensions of other social patterns (Co-authorship networks, …) Privacy and security