University of Konstanz
Graduiertenkolleg / PhD Program
Computer and Information Science

Spring Workshop 2010

Our DFG-sponsored PhD program at the University of Konstanz conducts a



date



location



program

08:15 Breakfast
09:00 Session 1: Talks - Chair: Mathias Heilig
Artistic Visualization of Seismic Data by Joonyong Ji
Real-Time Rendering of Forest Scenes by Soeren Pirk
Adapting Stress-Minimization for Dynamic Graph Drawing by Martin Mader
Analyzing Network Ensembles by Uwe Nagel
10:00 Session 1: Posters
10:30 Coffee
10:45 Session 2: Talks - Chair: Uwe Nagel
Advanced Visualization of Document Corpora by Hendrik Strobelt
Improvement Potentials in Attribute-based Opinion Analysis by Daniela Oelke
Processing Online News Streams for Large Scale Semantic Analysis by Miloš Krstajić
Analysis of Community-contributed Space and Time-referenced Data by Slava Kisilévich
11:45 Session 2: Posters
12:15 Lunch
13:00 Session 3: Talks - Chair: Martin Mader
The Search-Token - A Tool for Co-Located Collaborative Information Seeking by Mathias Heilig
PrIME: Primitives of Interaction for Multi-Display Environments by Mahsa Jenabi
Longitudinal Research in Human-Computer Interaction by Jens Gerken
ZOIL - A Post-WIMP User Interface Paradigm for Multi-Surface Interactive Spaces by Hans-Christian Jetter
Exploration and Visualization of Race Biking on Real Courses by Roman Byshko
14:00 Session 3: Posters
14:30 Coffee
14:45 Session 4: Talks - Chair: Mahsa Jenabi
CX-Factorisation for Spatio-Temporal Brain Activity Data by Martin Strauch
Multidimensional Classification of Dementia: Discriminant Analysis, Machine Learning and Visualization by Mirco Richter
The New Iris Data by Iris Adä
Teaching an XQuery Compiler to Handle Full-Text by Stefan Klinger
15:45 Session 4: Posters
16:15 Lecture: Dr. Leishi Zhang, University of Kent, UK: Let Pictures Talk
abstracts
17:00 End for students
Begin advisor's meeting


"Let Pictures Talk"

Part 1: Interactive Visualization for Dynamic Modelling of Gene Expression Time Series
Modelling temporal relations from time-course gene expression data has beenproven valuable in genomic disease diagnosis and genomic drug design. Because suitable software packages are lacking, we developed VisGene, a visual data mining system that facilitates the modelling of temporal gene regulatory networks based on gene expression time series. VisGene emphasises particularly the modelling of temporal relations. Since gene expression data are essentially high-dimensional, short Multivariate Time Series (MTS), the system reduces the dimensionality of the MTS before constructing a gene regulatory network using Dynamic Bayesian Network modelling methods. VisGene integrates the data mining algorithms in each stage of data analysis with effective visualization techniques to allow the investigation of the temporal relations between variables from different perspectives.

Part 2: Visualization with Euler Diagrams
Euler diagrams provide attractive visualization of overlapping information, because they represent intersection and enclosure of sets in an easily understandable way. They can be used in a wide range of application areas, such as visualizing statistical data, ontologies and the results of database search queries. However, practical use of Euler diagrams has been held back by difficulties in generating diagrams automatically. Until recently algorithms could generate diagrams only for a limited subset of abstract descriptions. In our project we developed general algorithms that, given any abstract description, can generate usable Euler diagrams. We proposed a number of additional layout improvement techniques. Furthermore, we developed algorithms for drawing Euler diagrams that meet additional constraints: for example, all contours are drawn as circles, or an algorithm for adding new contours to an existing Euler diagram without changing the original layout.