University of Konstanz
Graduiertenkolleg / PhD Program
Computer and Information Science

Graduation Talks

title

Visualization and exploration of financial time series data

speaker

Hartmut Ziegler, University Konstanz
Konstanz, Germany

date & place

Wednesday, 13.02.2008, 17:00 h
Room C252

abstract

The large amounts of data on the financial market today pose many computational challenges. Currently, companies like Reuters deliver around 50.000 data updates per second of financial stock market data. Such amounts of data can be analyzed by data mining algorithms, but for understanding market mechanisms and in order to evaluate the characteristics of assets, innovative visualization techniques in combination with traditional statistical techniques offer many advantages with regard to knowledge discovery in financial markets, as one image can communicate more insight than thousands of numbers. The talk will present research that helps to overcome the limitations of previous techniques, and demonstrate pixel-based visualization concepts by creating finger prints of financial assets which reflect the characteristics of an asset and significantly improve the insight into the behaviour of todays financial market. The work focuses on performance analysis of assets like visual analysis of growth rates over the full spectrum of all possible time intervals, inter-asset comparison of all assets on the market, visual pattern and trend recognition, portfolio analysis, visual exploration of long-term investments, weighting functions for emphasis on specific regions of interest, improved methods for performance and risk analysis, dominance plots and efficiency curves.