A Visualization Environment for Multiple Daytime Stock Price Predictions

Visualization of stock exchange market is one of the prominent applications of information visualization. Although there are several research results which visualize longterm stock price movements, visualization of daytime stock price movements is rare. However, it is a very important application domain, since the fact that there is no statistical method of predicting daytime stock prices indicates that the decision making process must be supported by tools for cognitive amplification. Even with the advent of agent-based simulation techniques, it is still a challenging issue because different parameters result in diverse predictions, and so the traders are forced to read multiple predictions quickly. In order to provide the simultaneous presentation of multiple predictions, we have built a system which supports line-charts with a cluttering detection and control mechanism, color-charts with Level-of-Detail control, and a workspace for overviewing the different prediction sets simultaneously. Although it is difficult to evaluate the system quantitatively, the system is novel in the sense that other commercial systems still use simultaneous display of windows each of which shows just a few charts. Since even trained traders cannot scrutinize more than ten windows, providing overviews of dozens of price changes is a very important function of the systems supporting financial traders.