Visualize Large Time Series with Interactive Temporal Aggregation


currently not available due to TU Wien budget cuts

Time series are important in many domains such as finance, environmental science, or medicine. Visual representations such as line plots are valuable tools for their exploration, but these time series often have more observations than fit on a computer screen. Examples: daily maximum temperatures over 110 years (40,419 observations), daily closing price of the IBM stock since 1962 (12,430 records), traffic on a highway ramp in 5 minutes intervals (50,400 records).

Data stream proposed by Bade et al. [2004] for high frequency dataTemporal aggregation (e.g., weekly averages) can be used to summarize time series and to provide an overview. Different visual representations are possible, for example box plots or the data stream proposed by Bade et al. [2004] (cp. Figure on the right). For effective interaction, the aggregation level should fit to the current zoom level and the transitions between aggregation levels can be animated. For focus and context displays (e.g., fisheye view, bifocal display), the aggregation levels can be different in focus area, possibly showing the raw data.

Specialized data structures for time-oriented data should be used to improve scalability for large time series.

Some features are optional depending on the scope and whether a group of students works on the project.


Develop visualization components for visualization of large time series with temporal aggregation based on the visualization toolkit prefuse and a preexisting toolkit for management of time-oriented data.  Demonstrate your results in a prototype, test scalability improvements with large time series, and/or perform a user study.

Other information

Bade, Ragnar, Stefan Schlechtweg, and Silvia Miksch [2004]. Connecting Time-Oriented Data and Information to a Coherent Interactive Visualization. In E. Dykstra-Erickson and M. Tscheligi (Eds.), Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, (pp. 105–112).

Information about prefuse:


Further information

Time series, optional: Medical Informatics, Business Informatics
Visual Analytics (VA)
Previous knowledge
Java, optional: prefuse, user studies