Oracle Usable Apps | Applications User Experience Simplicity, mobility, extensibility
Customer Role: How User Experience Research Is Conducted
Correlate and Compare Graphs

The Correlate and Compare Graphs patterns and best practices help project managers, developers, and designers select visualization techniques such as tables and graphs, and helps users to understand how multiple graphs are used together and to perform tasks related to comparing and correlating information.

The following information was originally written for Oracle Fusion Applications. To closely align these best practices with Oracle Business Intelligence Enterprise Edition (OBIEE) 10g, additional examples from OBIEE have been included here.
Using Multiple Graphs for Correlating and Comparing Graphs
When you need to display multiple graphs, present them in a way that enables the user to compare and consolidate information efficiently. The following sections discuss several possible ways of doing this.
Single Graphs Return to Top

Graphs can take a lot of screen space. Before you decide to use multiple graphs, make sure that a single graph will not suffice. A single graph may be faster and easier to understand than multiple graphs because users need to move their eyes back and forth fewer times to compare multiple data sets.

Use Line, Clustered, and Stacked Bar Graphs

To determine if a single graph will meet your needs, review the high-level tasks that the user is trying to accomplish. Whatever graph you choose must support the user's primary task, and, if possible, the user's secondary task as well. If the data that supports the primary and secondary tasks is in the same unit of measure, plot both series in a single graph. A simple example would be a comparison of forecast and actual expenses over a common time period, which could be shown as two lines on a single line graph. For a large number of series that a user may need to compare, consider providing parameters to enable users to show only the series that they want to compare. For example, quarterly sales figures for several products over a common time period could be shown as a clustered bar graph.

Use Dual-Y and Split-Y Graphs

For tasks that require comparing data sets with different units or scales, dual-y graphs may be appropriate. Dual-y graphs use two vertical scales, one on each side of the graph, but share a common horizontal scale. If possible, use the same color for the plotted data and the associated scale.


Figure 1: Dual-Y Graph


Dual-y and split-y graphs can be used to compare data in different scales or units of measure.

Use Combination Graphs

You can also use a combination graph, which allows combinations of areas, lines, and bars. Like dual-Y graphs, combination graphs typically have a separate axis for different data sets.


Figure 2: Combination Graph

To Show Only a Difference, Consider Transforming the Data
If the user's primary task is to appreciate a difference between two data sets, consider plotting the difference. For example, if the primary task is to appreciate the difference of forecast vs. actual expenses, plotting the difference may be more useful than plotting both the forecast and the actual expenses, and expecting the user to visually subtract them.
Two Series

Users may more easily appreciate a difference between two series (left) when the difference is calculated and plotted (right).

Multiple Graphs: Graph Sets Return to Top
Applications with multiple graphs can be complex to use, so we strongly recommend that you take the time to identify the users' most likely tasks and identify the best possible graphs to support those tasks. On the other hand, if you are designing a complex work area to support multiple tasks, you should attempt to establish a mapping among graph sets that is most appropriate for helping users accomplish their most likely tasks.

In advanced analytical applications, you may not able to predict the most appropriate graphs to present. You may have to present users with the option to select the graphs that they want to see in a secondary window or dynamic tab.

Graph Sets Supporting a Single Task Return to Top

Embed Adjacent Graphs Directly in the Page

You should use adjacent graphs and avoid using a drop-down menu to switch between graphs whenever possible. Using a drop-down menu to switch between graphs saves screen space, but it makes comparing graphs almost impossible because the user can see only one graph at a time. As an alternative, review your layout to see if the screen real estate allows enough room for adjacent graphs.

Embed Adjacent Graphs in a Tab

If you don't have enough space to embed the graphs directly in the page, put the graphs on an auxiliary tab labeled "Analytics." While the use of a tab is not as optimal as embedding the graphs directly in the context of your transaction, at least users can still easily access the graphs and compared them side by side.

Graph Sets Supporting Multiple Tasks Return to Top

When you need to support multiple tasks, you need a way to switch between multiple associated graph sets.

Switch Graph Sets by Selecting Table Rows

One straightforward way to update graph sets is to use row selection on a table in which each row is associated with a task.

When graph sets get large, you should organize them into sub-groups. In the previous example, you could organize the graph set for helping a user manage a back-ordered item task in an Order Fulfillment flow of a Supply Chain Management application into three sub-groups that show information about:

  • The back-ordered item
  • The associated order
  • The associated customer

Within each group, you should display the graph that the user will most likely find useful first. Use a drop-down menu to enable users to access the other graphs in each group.

General Guidelines for Using Multiple Adjacent Graphs Return to Top
Share a Scale or a Legend If Possible

Sharing a scale or a legend saves screen space. Users find it easier to compare values that use the same scale and to compare graphs that use the same mapping of colors to categories. Of course, graphs can only share a single scale if they show quantities in the same range and units, and they can only share a legend if they show the same set of categories. If you need multiple gauges, consider using a gauge set, which is designed to allow multiple gauges to share a single scale and legend.

Support Comparisons of Similarly Transformed Data

Presenting data in a form that supports comparison can be misleading if the data has not been collected, measured, or transformed in a consistent manner. When users need to compare data that has been collected over time periods of different lengths, be sure to label the differences clearly.

Use the Same Graph Types If Possible
Users find it easier to compare graphs of the same type than graphs that are of different types. For example, instead of using a bar graph and a pie chart, consider using two bar graphs.

Check That Colors and Other Settings Are Used Consistently

While most graph settings are determined automatically for OBIEE applications, extra care is needed to ensure that the default settings for individual graphs minimize meaningless visual differences between multiple graphs. Users comparing graphs expect that visual differences will be significant. For example, if one graph uses a certain color to display a particular category, make sure that category is displayed with the same color in every other graph on the page. Check that similar colors are used to indicate similar categories.
You many need to override other settings for graphs that display different amounts of data. For example, when using two pie charts , users may be confused if one pie has labeled slices, while another does not. Users may also be confused if the pies themselves are different sizes.  |  About Oracle  |  Careers  |  Contact Us  |  Legal Notices  |  Terms of Use  |  Your Privacy Rights