19th International CODATA Conference
Category: Data Visualization

VidaMine: User-Centred Development of a Visual Mining Environment

Stephen Kimani (kimani@dis.uniroma1.it), University of Rome "La Sapienza", DIS, Italy
Stefano Lodi (slodi@deis.unibo.it), University of Bologna, DEIS, Italy
Tiziana Catarci (catarci@dis.uniroma1.it), University of Rome "La Sapienza", DIS, Italy
Giuseppe Santucci (santucci@dis.uniroma1.it), University of Rome "La Sapienza", DIS, Italy
Claudio Sartori (csartori@deis.unibo.it), University of Bologna, DEIS, Italy


Tremendous technological breakthroughs have virtually revolutionized the world. One major consequence is that humans are confronted with ever-increasing already massive amounts of data at virtually every turn. On the other hand, there have not been corresponding advances in techniques for extracting knowledge from the data. It therefore comes as no surprise that data still present formidable challenges to effective and efficient mining of knowledge.

Since the human-visual system enables both recognition and understanding of overwhelming data at an instant [14], it is an outstanding resource for detecting and extracting knowledge from data. Tapping into the human-visual system would primarily entail exploiting relevant and effective visual strategies within the user interface. Most mining efforts have employed such visual strategies only at the beginning and at the end of the discovery process [13]. Human involvement in the entire mining process is crucial. Toward that, a human-user architectural component should be designed and positioned at a strategic place in an open overall discovery framework. Such an approach constitutes a great step toward according the user a central place in the entire discovery process since the aforementioned human user's outstanding visual system becomes much more available for exploitation across all the phases of the discovery process.

This research, which is part of the project D2I (Data to Information, http://www.dis.uniroma1.it/~lembo/D2I), focuses on the investigation and exploitation of strategies that are instrumental toward the realization of a visual interaction environment that supports the human user throughout the entire process of mining knowledge. Our research findings culminated in the realization of VidaMine (VIsual DAta MINing Environment) [1--12], a visual data mining system that exploits various visual strategies thereby offering a visual interface that allows or enables the user not only to process data, but also to steer, guide or direct the entire process of data mining.

Besides involving the user in the entire mining process, a visual mining system ought to involve the users in the user interface design process; which is not common in many visual mining endeavors. We adopted a real user-centered user interface design, equipped with usability studies. As reported in [3], we employed various usability methods progressively in the development lifecycle.

Moreover, unlike with many visual mining efforts and as reported in [4, 12], VidaMine is developed based on a careful definition of visual syntax and formal semantics. Among other benefits, such a definition facilitates data exchange and capturing semantics.

References:
1. S. Kimani, S. Lodi, T. Catarci, G. Santucci and C. Sartori: "VidaMine:A Visual Data Mining Environment". Journal of Visual Languages and Computing 15 (1):37-67, Elsevier, 2004.
2. S. Kimani, T. Catarci and G. Santucci: "A Visual Data Mining Environment". Visual Data Mining: Theory and Applications, S.J. Simoff, M. Noirhomme-Fraiture, M.H. Böhlen Ed.s, LNAI series, Springer-Verlag (to appear).
3. S. Kimani, T. Catarci and G. Santucci: "Visual Data Mining: An Experience with the Users". Proceedings of HCI International - Universal Access in HCI: Inclusive Design in the Information Society, 2003.
4. S. Kimani,
S. Lodi, T. Catarci, G. Santucci and C. Sartori: "Visual Data Mining with VidaMine". Proceedings of the Italian Symposium on Advanced Database Systems (SEBD), 2003.
5. S. Kimani: "An Effective Visual Data Mining Environment". Doctoral Posters of the International Conference on Very Large Data Bases (VLDB), 2002.
6. S. Kimani, T. Catarci and G. Santucci:  "A Visual Data Mining Environment: Metaqueries and Association Rules". Proceedings of the International Conference in Advanced Visual Interfaces (AVI), 2002.
7. F. Angiulli, T. Catarci, P. Ciaccia, G. Ianni, S. Kimani, S. Lodi, M. Patella, G. Santucci and C. Sartori: "An Integrated Data Mining and Data Presentation Tool". Proceedings of the International Conference on Data Mining Methods anf Databases for Engineering, Finance and Other Fields, 2002.
8. S. Kimani, T. Catarci and G. Santucci: "A Visual Data Mining Environment". Proceedings of the CODATA Workshop on Information Visualization Presentation and Design, 2002.
9. S. Kimani, T. Catarci and G. Santucci: "A Visual Data Mining Environment". Proceedings of the ECML/PKDD Workshop on Visual Data Mining, 2002.
10. S. Kimani, T. Catarci and G. Santucci: "Visual Data Mining. Il Sistema VidaMine". Proceedings of the D2I Workshop on Integration, Warehousing and Mining of Data from Heterogeneous Sources, 2003.
11. T. Catarci, P. Ciaccia, V. Curci, S. Kimani, G. Ianni, S. Lodi, L. Palopoli, M. Patella, G. Santucci and C. Sartori. "Visual Data Mining System Architecture". Technical Report D3.R2 of D2I, Integration, Warehousing, and Mining of Heterogeneous Data Sources, Italian MIUR Project, http://www.dis.uniroma1.it/~lembo/D2I/, 2001.
12. S. Kimani, S. Lodi, T. Catarci, G. Santucci and C. Sartori. "VidaMine: A Visual Data Mining Environment". Technical Report, IEIIT-CNR, Bologna, Italy, September 2003.
13. U. Fayyad, G. G. Grinstein and A. Wierse. "Information Visualization in Data Mining and Knowledge Discovery". Morgan Kaufmann Publishers, 2002.
14. S. K. Card, J. D. Mackinlay and B. Shneiderman. "
Readings in Information Visualization-Using Vision to Think". Morgan Kaufmann Publishers,1999.]