Applying Animation to the Visual Analysis of Financial Time-Dependent Data

From InfoVis:Wiki
Revision as of 22:59, 8 May 2008 by UE-InfoVis08-07 (talk | contribs) (added citations, data types, evaluation)
Jump to navigation Jump to search

Authors

Abstract

For decades, financial analysts have strived to use modern data visualization tools to improve the timeliness and quality of their analysis. As the amount of data to be processed increases rapidly and requirements on quality of financial analysis rise, the demand for analysis support systems grows. We present a system for the visual analysis of large amounts of time-dependent data using animation. For each data entity, indicators are presented in a scatter-plot framework, displaying the correlation between them. The design of the glyphs illustrates additional data dimensions. The system uses animation to handle the time-dimension of the data. It offers various features, such as focus, zoom, details on demand and time period selection to support the analysis. Financial indicators are used to demonstrate the usability of the system. The animation proves to be a powerful tool for analysing time-dependent processes in cross-sectional data sets and discovering patterns in the data.
[Tekusova and Kohlhammer, 2007]


Figures


The figure shows the relationship between return and volatility for some example asset categories. It is usual that riskier assets can mean higher returns and vice versa.

The relationship between volatility and return over time is at the centre of the investors' atteintion.
[Tekusova and Kohlhammer, 2007]





The FinMotion userinterface is divided into two areas:

  • Main visualisation window
  • Interaction panel

The window contains five panes as it can be seen on the picture.

  • Pane 1: "Scatter plot view": Is used for viewing scatter plots which are very important for the analysis. The interaction panel consists of tools to manipulate the scatter plot view.
  • Pane 2: "Data": Provides tools to manipulate data and get data descriptions.
  • Pane 3: "Data selection": Provides features for selecting data.
  • Pane 4: "View settings": Provides more features for data visualization and manipulation.
  • Pane 5: "Animation": Provides tools for manipulationg motion features.




Negativ values in the scatter plot are represented as unfilled rectangles. The size of the rectangle represants the absolute value of the data point.


The animation of data points can be a very powerful tool but it hides the history. Therefore users can choose to show trajectory paths during animation.


Focus and Zoom: Users can define "target stocks". Selected data items keep their original colour and other items are coloured in grey. This gives the users the possibility to concentrate on interesting data. This feature is also very useful because of the overlapping glyphs problem.


This picture shows the "Details on Demand" feature of FinMotion.

As users may wish more information on visualised data, a legend and further information on demand are provided.
[Tekusova and Kohlhammer, 2007]





The legend text placement was a major challenge. The legend text can be places somewhere around the glyph in the grey area. The algorithm the authors used works fine, but it is impossible to overcome the overlapping legend problem completely.

Important citations

The usability study showed that the animation can reveal new unexpected patterns in the data which cannot be perceived in a static analysis of correlation. Furthermore, animated data visualization better reflects real market movements.
[Tekusova and Kohlhammer, 2007]


Data types

FinMotion is suitable for daily stock market data or any other kind of financial data with temporal information.

Evaluation

A usability survey has been carried out and the feedback was very positive. The software has shown an incredible ease of use and the animation feature was also very positively ranked. Some new features have been added upon request.