Using Topic Maps for Visually Exploring Various Data Sources in a - PowerPoint PPT Presentation

Using Topic Maps for Visually Exploring Various Data Sources in a Web-based Environment � Nadeem Bhatti, Fraunhofer IGD Eicke Godehardt, SAP Research � TMRA 07: 2007-10-11

Overview Introduction � Scalability � Prefeching � Garbage Collection � Data Sources � Technologies for the rich interactive web applications � Case Study � SAP-TM-Viewer � Conclusion � Archivierungsangaben 2

Introduction � Business need � Huge amount of data � Complex dependencies � Heterogeneous data sources � Challenges � Intuitive and user friendly graphical user interface � Visualizing large amount of data from already existing data sources � Easy navigation/exploration through the data � Web-based and easy integration in existing websites (Web 2.0, Mashup) � Approach � Graphical metaphor for topic maps � Project Partner Archivierungsangaben � Fraunhofer IGD and SAP Research 3

Scalability � Challenges � Huge amount of data (10K to 40K) � Complex dependencies � Web Solution � Visualizing large amount of data Archivierungsangaben 4

Scalability – Loading on Demand Normal course: � users selects a topic 1. context is loaded 2. merged with local topic map 3. and displayed 4. Internal status of topic map: Archivierungsangaben 5

Scalability – Pre-fetching � Normal course: 1. users selects a topic 2. direct direct context is loaded 2. 3. merged with local topic map 4. and displayed 5. next level of context is loaded next level of context is loaded 5. 6. go to 4 go to 4 6. � Smoother experience Internal status of topic map: Archivierungsangaben 6

Scalability – Garbage Collection � Remove unnecessary topics from local Topic Map � After some time of using/browsing: � big internal topic map � multiple cluster/island of data (not accessible by browsing) � may slow down the viewer � Possible criteria: � how long ago is the last time of display � how far away is a topic from current topic � Justification � no server side changes are propagated Archivierungsangaben 7

Archivierungsangaben Data Sources 8

Technologies for the rich interactive web applications (I) � We compare different possible technologies: � Flash™ � JAVA™ Applet � other (AJAX, JavaScript …) � Indicators: � Performance � Visual effects (User Experience Design) � Ease of distribution Archivierungsangaben 9

Technologies for the rich interactive web applications (I) � Flash™ Flash™ JAVA JAVA™ Applet ™ Applet � + Availability/Penetration (over 98%) � - Availability (<85%) � + Rich interactive web applications � - Rich interactive applications (Designer/SW-Developer) (Designer/SW-Developer) � + Easy XML loading/handling � + Libraries � +/- � +/- Archivierungsangaben 10

Use case/Case study � SAP-TM-Viewer � Challenges � SAP applications integrated information system � provide the context specific help � huge amount of knowledge items � visualize relationship and hierarchical structure of knowledge items � SAP business data Model � Work Center ▪ Work Center View � Business Scenario ▪ Business Scenario Variant � Business Configuration ▪ Business Package Business Topic ▪ Archivierungsangaben � Enterprise SOA 11

Use case/Case study History Association Cluster Topics Sectors Levels Archivierungsangaben 12

Conclusion and Future Work � Using Topic Maps for Visually Exploration � Scalability � Data sources � Technologies for rich interactive web applications � Case Study � SAP TM-Viewer � Future research topics � Collaborative topic maps evolution � Awareness (visualize collaborative changes) � Aspect oriented topic map views Archivierungsangaben 13

� Thank you! Thank you! � Nadeem Bhatti deem Bhatti Research As Research Associate sociate Fraunhofer IGD Fraunhofer IGD � Fraunhoferstraße Fraunhoferstraße 5 64283 Darmstadt 64283 Dar stadt, Ger , Germany any � nadeem.bhatti@igd.fraunhofer.de nadeem.bhatti@igd.fraunhofer.de � Several post-doc Several post-doc, PhD PhD candidate and student candidate and student Archivierungsangaben positions open in Darmstadt positions open in Darmstadt 14