Computer graphics processing and selective visual display system – Computer graphics processing – Graph generating
Reexamination Certificate
2000-07-13
2002-08-06
Jankus, Almis R. (Department: 2672)
Computer graphics processing and selective visual display system
Computer graphics processing
Graph generating
Reexamination Certificate
active
06429868
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to the display and analysis of quantitative data. More particularly, the invention relates to a method and computer program for simultaneously displaying very large sets of quantitative data without first standardizing or conditioning the data, while still permitting viewers to easily distinguish individual variables and values in the display, and allowing wide control over a color mapping process.
2. Description of the Prior Art
The ever-increasing use of computers has expanded the amount of quantitative data available for analysis. For example, both professional and amateur stock traders now have access to tremendous amounts of data that can be used to track and analyze stocks, and health care providers can monitor a plurality of vital statistics within an intensive care environment. As computers become more ubiquitous, the amount of quantitative data available for analysis is expected to continue to grow at an extremely rapid rate.
Quantitative data is often most useful when it is displayed so that viewers can see or visualize trends or patterns in the data. For example, stock traders often desire to view time series of several stock prices and volumes and to compare the performance of the stocks, or compare entire sectors of stocks made up of hundreds of time series.
Many different types of graphs and data mapping techniques can be used to display quantitative data for analysis purposes including line graphs, bar graphs, area graphs, surface graphs, two, three and four-dimensional contour graphs, bubble graphs, column graphs, heatmaps, treemaps, etc. Many of these techniques can also be combined with color mapping procedures wherein data values within the graph or data map are indicated through means of a color display or color “values”. Color can be used to indicate a value and also to enhance certain characteristics of the data, or to indicate priorities or alerts. Color mapping is used, for example, in line graphs, contour maps, heatmaps, treemaps, and in imaging applications including medical imaging, radar, and other sensor data display.
Unfortunately, these prior art graphs, data displaying, and color mapping techniques suffer from several limitations to their utility. One limitation is that prior art graphs are generally limited in the number of variables which they can display simultaneously while still allowing differentiation of individual variables and/or values within the graph. For example, a line graph with six or more time series or variables typically looks cluttered and the data therein becomes intertwined, making differentiation between individual time series variables difficult if not impossible. One solution is to stack multiple graphs to view simultaneously, but this solution is limited in the number of graphs that can be displayed and it can be difficult to make comparisons across the graphs.
Another limitation of prior art graphing and data displaying techniques is that for a number of series to be effectively graphed the data needs to be in a relatively narrow or common range. This problem occurs both in line graphs and in color contour mapping. On a line graph, when the data is not in a common range, it must be transformed, normalized, or standardized to a common scale, or a variable must be selected which is in a common range. If not in a common range some of the time series may be difficult to distinguish and appear no different from zero. For example, if ten stock prices, a market index, and a market volume are to be graphed together and the stock prices have values ranging between $4 and $250, the market index has values ranging between 5,000 and 10,000, and market volume has values between 500 million and 2 billion, on a common graph scale the time series with the lower values become indistinguishable from zero. One solution to this problem is to provide separate Y-axes, however, this solution is limited to graphs containing only a few variables. In many cases the observer desires to see as many variables simultaneously as possible. Another solution is to mathematically transform or normalize the data to a common range, for example, by taking the logarithm of the data, standardizing the data, mapping the raw data to a relative index based on a reference point or to some other variable with a common range, for example percentage change. This solution is limited because observers may have difficulty inverting the transformation to determine the actual value of the raw data which may be of interest, and also the need to choose variables with a common range severely limits the choice of data. In addition, data transforming, conditioning, or normalizing for graphing or other displays can be a laborious effort and often needs to be done on a case by case basis, the approach to use depending on the type and particulars of the data.
In a limitation related to time series being indistinguishable from zero on a line graph, treemaps are displayed with an “area-coded” variable which determines the size of a rectangle displayed on the screen, small area coded variables are difficult to distinguish or find on the treemap graph, the analog of being indistinguishable from zero on a line graph. Treemaps can be rotated to view from different angles but the process does not guarantee that small area-coded variables will be found, and the correct angle for viewing is uncertain.
A similar limitation related to scale exists in color mapping of data using contour color mapping approaches as found in two, three, and four dimensional contour maps, in heatmap and treemap applications, and in imaging. Contour color mapping uses the entire data space or data matrix (image) as the basis for the color process. Unless the data is in a common range, it may happen that only the most extreme colors in a color spectrum made up of a number of colors ordered from high to low, will be used. In the example of stock prices, market index, and market volume, mapped on the same contour graph or on a heatmap or treemap, the volume will use only the highest color and the small stock prices use only the lowest color. The color mapping process loses all its details. The prior art solutions to this problem are generally the same as used on line graphs. Mathematical transformations to a common range are used or the choice of variables is limited to those in a common range. These approaches suffer from the same limitations as the line graph solution, it is difficult for the observer to invert the transformed value to relate to the raw data value, and restricting the choice of variables to those in a common range severely restricts the utility of the approach.
Another limitation related to scale occurs with long and trending time series. Displayed on a line graph, when a narrow area (domain of time) of the graph of the trending series is viewed, with the scale or Y-axis set for the full data set, the area viewed appears flat. A similar problem happens in color mapping, when a small domain of the time series is viewed with the color process based on the full time series domain, the part viewed utilizes only a narrow band or few colors of the color spectrum, the analog to appearing flat on a line graph.
A related limitation is the effect of outlier data points on color mapping. An outlier is a value within a data set that is significantly different from the range of the rest of the data, e.g., beyond plus or minus 3 standard deviations. Outliers affect the color mapping by “absorbing” many colors. That is, the outlier will be assigned the highest (lowest) color in an ordered color spectrum, and there will be many colors unused between the outlier and the rest of the data, and then leaving relatively few colors to differentiate the range where most of the data is located. One solution is to provide methods to remove the outliers. Another solution is to index the outliers under a mathematical transformation. However these solutions are limited in their scope and capability.
Another limitation of prior art graphing
Dehner, Jr. Charles V.
Hewitt Wayne
Cunningham T. F.
Hovey Williams Timmons & Collins
Jankus Almis R.
LandOfFree
Method and computer program for displaying quantitative data does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and computer program for displaying quantitative data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and computer program for displaying quantitative data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2942661