Compact Display of Event Episodes in Multiple Time-Series

Problem

"We are addressing the problem of designing a compact display for multiple event-based long time series data, which would provide global overview of the whole dataset and allow atomic analysis of individual events, i.e. time-stamped data points that arrive in data streams at non-equidistant time intervals. Additionally, each event contains not just a numerical value, but is rich in content, providing meta information that is of great interest for the user. Therefore, it is important to allow direct interaction with events on the atomic level. Events can belong to different categories, or we can acquire different event data streams for which simultaneous analysis is needed." [Krstajic et al., 2011] 

Aim

- develop a prototypical implementation of the Cloudlines visualization technique;

- (only for theses) design and implement specific adaptations for medical data, e.g. clinical actions performed on large cohorts of patients.

Other information

Krstajic, M.; Bertini, E.; Keim, D.;  CloudLines: Compact Display of Event Episodes in Multiple Time-Series, IEEE Transactions on Visualization and Computer Graphics, vol.17, no.12, pp.2432-2439, Dec. 2011

http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6065010&isnumber=6064926

Contact

Further information

Topics
event based data, e-health
Area
Information Visualization (IV)
Previous knowledge
Java (preferably Swing, Java2D, prefuse Toolkit)
English
Scope
BA
PR
MA
Assigned as
Bachelor thesis/Bakkalaureatsarbeit
Student(s)

Sascha Rainer

Stefan Schlossarek

Status
closed