Browse Prior Art Database

Method to filter least significant data from client rendered charts for performance while maintaining chart integrity

IP.com Disclosure Number: IPCOM000031134D
Original Publication Date: 2004-Sep-14
Included in the Prior Art Database: 2004-Sep-14
Document File: 4 page(s) / 47K

Publishing Venue

IBM

Abstract

This article describes a method to reduce the size of a server generated document representing a client rendered chart. The method decreases the document size by removing data points which do not provide significant information about changes in the data. Benefits of doing so are decreased transfer time over the network, decreased time to render charts on the client and a visually clearer and more responsive chart.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 4

Method to filter least significant data from client rendered charts for performance while maintaining chart integrity

Disclosed is a method to reduce the size of server generated chart documents in order to boost client response time. With the demand for highly interactive web based applications, new technologies which provide interactive graphical representations of data are becoming more popular. Examples of these technologies include Java* applets and SVG (Scalable Vector Graphics). Because these technologies can render data charts on the client side, they also open up the possibility for complex user interactions at the client side. Usually the data that is rendered is downloaded from a server and is often in XML document format. Unlike static image charts which have a fairly fixed size that is proprotional to the chart's dimensions on the screen, the size of these data documents is linearly proportional to the amount of data contained with them. In cases where a large set of data is contained, the data documents can become lengthy and can take significant time to download and render. They can also make interaction on the client side slow and frustrating.

The solution is to provide the user with a mechanism to reduce the amount of data shown on a chart while minimally impacting the value of the chart to the user. On the client the user will have access to filtering *controls that then convey the user's intent to the server on subsequent requests for chart data. It is the server that will use an intelligent algorithm to reduce the amount of data that is returned to the client. When the chart is rendered with the reduced data, the important fluctuations in the data will still be present. Closely spaced points causing minor or no fluctuation will have been removed.

Figure 1 below shows a representation of a chart containing a large amount of data. Figure 2 shows a chart in which least significant points have been removed.

Figure 1: a chart containing a large number of data points

1

[This page contains 2 pictures or other non-text objects]

Page 2 of 4

Figure 2: a chart displaying the data after least significant points have been removed

Filtering is configurable by the user. At a minimum the user can choose to turn filtering on or off. A more advanced solution allows the user to choose from several degrees of filtering or even allows them to adjust individual filtering parameters. This configuration is done on the client before or during the viewing of a chart. The client uses standard input devices such as buttons, combo boxes, and text entry fields to gather preferences and would then relay the preferences to the server on requests for chart data. In a web environment, this information would most likely be relayed via the HTTP request parameters.

Once the server receives a request for a chart along with filtering preferences, it is up to the filtering implementation to determine how filtering should be done. There are certainly...