4 Using the Dashboard

Specifying the Window of Time to Display

To specify the window of time that you want the event window to present, click one of the four time values at the top of the page. By default, you are presented with increments of 60 minutes, 45 minutes, 30 minutes, or 15 minutes. You can adjust the increments displayed on the page by clicking the Change Time Options button and then selecting a new set of time values. Available time windows from which you can select are the following:

• 24 hours, 18 hours, 12 hours, 6 hours

• 4 hours, 3 hours, 2 hours, 1 hour

• 60 minutes, 45 minutes, 30 minutes, 15 minutes

• 10 minutes, 8 minutes, 5 minutes, 2 minutes

Specifying the Composite Event to Display

To change the composite event displayed in the Real-Time Viewer:

1. In the Composite Event box, select the composite event you want to display. Note that you can select standalone composite events (those not part of a composite event group) or composite event groups. However, composite events that are part of a composite event group are not presented in this box.

   Standalone composite events and composite event groups are presented in the Real-Time Viewer as composite events. Composite events that are part of a composite event group are not displayed separately from the group.

2. Click Apply.

The Real-Time Viewer refreshes to display data for the selected composite event.

Specifying the Dimensions to Display

Using the Dimension control, you can specify that composite event instances be displayed along a dimension.

To specify the dimension to display:

1. In the Dimension box, select the dimension you want to display.

2. Click Apply.

The chart refreshes to present instances of the selected composite event along the selected dimension.

For example, using the loan flow example presented in Figure 1-1, suppose the administrator defined these two dimensions: CarMake and CarModel for auto loans. You can specify one of these dimensions as the category by which to present composite events in the Real-Time Viewer. Figure 4-5 shows the Real-Time Viewer when CarMake is selected as the dimension by which to categorize composite events; Figure 4-6 shows the Real-Time Viewer when CarModel is selected as the dimension by which to categorize composite events.

Specifying the Type of Composite Events to Display

To specify the type of composite events that are included in the Real-Time Viewer:

1. In the Events box, select the type of composite event you want to display.

   The text following this list describes the types from which you can choose.

2. Click Apply.

The Real-Time Viewer refreshes to display the selected type of composite events.

The types of composite events from which you can choose are as follows:

• All

  When you select All from the Events box, all event instances captured by Oracle BPEL Process Analytics are presented in the Real-Time Viewer (chart and table).

  Figure 4-7 shows an example of a chart that contains all returned event instances, including timed-out events. Each marker (circle, diamond, square, plus sign, and x) represents a different event instance within each composite event instance. Line segments connect event instances that are part of the same composite event instance.

• Complete

  When you select Complete from the Events box, only composite event instances for which end events have been returned are presented in the Real-Time Viewer. Figure 4-8 shows an example of an chart that contains only complete composite events.

• Pending

  When you select Pending from the Events box, only composite event instances for which an end event has not yet been returned (and which are not timed out) are presented in the Real-Time Viewer. Figure 4-9 shows an example of an Real-Time Viewer that contains only pending composite events.

• Timed Out

  When you select Timed Out from the Events box, only timed out composite event instances are presented in the Real-Time Viewer. A composite event instance is considered to be timed out if the next event in the process flow represented by the composite event has not been received within the time period specified by the time-out value. The time out value is specified by the administrator when the composite event is defined.

  A chart containing only timed-out events looks similar to an chart containing only pending events.

Scrolling and Refreshing the Real-Time Viewer Chart Data

The left arrow and right arrow under the chart allow you to scroll back and forth through the chart data. Click the left arrow, to scroll back one increment of time. Click the left arrow several times to scroll back additional increments of time. The time increment is one-half of the time unit you selected at the top of the page, as described in "Specifying the Window of Time to Display". Similarly, click the right arrow, to scroll forward a unit of time.

To refresh the chart data and resume the presentation of data for the current time window, click Real-Time Refresh. By default, the chart is refreshed automatically every minute.

The View Mode is specified as Real-Time when you are viewing the current time window; it is specified as historical when you are viewing any other time window.

Showing and Hiding Connectors

In addition to specifying which composite events are included, you can specify whether or not connectors between related events within a composite event instance are connected. Click Hide connectors to hide the connectors; click Show connectors to show the connectors.

Figure 4-10 shows the Real-Time Viewer with the connectors hidden. Figure 4-8 shows the Real-Time Viewer with the connectors displayed.

When the events within a composite event occur within a relatively short time period, it can be useful to include the connectors between events. When the events within a composite event occur over a long period of time, days for instance, then the connectors will probably be of little value, and you might want to hide them.

Viewing Event Data in a Table

To view the chart data in a tabular format, click View Event Data. A page opens with a table such as that shown in Figure 4-11, which presents the dimension name, if applicable, the time at which the event occurred, the type of event, and the eyeglass icon, which is a link to access composite event details.

See "Accessing Composite Event Details from the Chart or Event Data Table" for information about displaying composite event details.

Accessing Composite Event Details from the Chart or Event Data Table

You can access composite event details from both the chart and the Event Data table, as follows:

• In the chart, click a data point within the chart.

• In the Event Data table, click an Event Type name or the eyeglass icon.

In all cases, the Composite Event Details page opens. However, when you click a data point in the chart or an Event Type name in the Event Data table, the Composite Event Details page opens with the navigation tree expanded to display the attributes of the selected event type, as shown in Figure 4-12. When you click the eyeglass icon, the Composite Event Details page opens with the navigation tree collapsed, as shown in Figure 4-13.

The Composite Event Details page presents the composite event in a navigation tree. You can expand the navigation tree to view the events and the attributes of the events within the named composite event. For each event, the time at which the event occurred is presented. For each event attribute, the attribute type and value are presented.

Viewing and Understanding Radar Charts

A radar chart displays KPI values in relation to their performance bands. Performance bands are defined by the Oracle BPEL Process Analytics administrator. A radar chart can display either multiple KPIs or multiple dimensions of one KPI.

When performance bands are defined, they allow a KPI to be divided into discrete bands to categorize performance. Oracle BPEL Process Analytics supports five performance bands, based on five performance categories relative to a target value. (For example, the default categories for performance bands are 0 - 20%, 20% - 40%, 40% - 60%, 60% - 80%, and 80% - 100%). These bands are also referred to as weak, under, normal, over, and strong performers, respectively.

Reading a radar chart is best explained using an example. Figure 4-15 shows a sample radar char. This figure presents the sales performance of five KPIs over 2 days. A blue polygon represents KPI performance relative to target values on March 16th and a green polygon represents KPI performance relative to target values on March 17th. The radar chart shows:

• AvgLoanAPR was a normal performer on both days.

• CountLoanOffers was a strong performer on the 16th, but an under performer on the 17th.

• CountLoanOffers4BadCredit was a strong performer on the 16th, but an under performer on the 17th.

• CountReq was a strong performer on the 16th, but an under performer on the 17th

• CountReqReject was a strong performer on the 16th, but a normal performer on the 17th.

As shown in Figure 4-15, due to limited space, KPI labels can be cropped. To view the full KPI name, place the mouse cursor over the KPI label; a pop-up window displays the full KPI name. Also note that radar charts are optimally displayed when the target screen size is set to 800 x 600 pixels. At other sizes, the radar chart appears off-center. The Oracle BPEL Process Analytics administrator can reset the target screen size, as described in "Setting the Dashboard Screen Size".

To drill down in the radar chart and view KPI detail pages, click a KPI (or KPI dimension) label within the chart.

Accessing the Real-Time Viewer for a KPI Presented in a View Pane

After viewing a KPI in a view pane, you may decide you want to see the real-time delivery of the event instances within the composite event on which the KPI is defined.

To access the Real-Time Viewer from a multiple pane view, click the Real-Time Viewer icon (shown in Figure 4-4) that appears in the upper-right corner of the KPI view pane for which you want to view events.

If a multiple pane view contains data from multiple composite events, then multiple Real-Time Viewer icons will be presented. Moving the cursor over any one of the Real-Time Viewer icons causes a pop-up window to open and display the composite event that will be displayed in the Real-Time Viewer if you click that icon.

Viewing and Understanding KPI Details

KPI detail pages enable you to analyze, in depth, a KPI presented in a multiple pane view. Four KPI detail pages are available, as described in the Table 4-1.

Viewing and Understanding Performance Details

The Performance Detail page presents KPI performance data over time. If the administrator has specified target values for a KPI, the target values are also presented, so that you can compare the actual performance values to the target performance values.

To access performance details:

1. In the Dashboard, select a multiple pane view from the View box.

2. In the multiple pane view, click a data point in any chart or table. A Performance Detail page opens.

3. In the Switch KPI box, select the KPI for which you want to view performance. The Performance Detail page opens for the selected KPI.

Controls in the Performance Detail page allow you to specify the dimensions by which KPIs are filtered, the specific dimension values to present, the type of chart in which to present the data, and how far back from the current time you want the chart to include data. As with charts in multiple pane views, you can determine the exact value of an element in a chart presented in the Performance Detail page by placing the mouse cursor over the chart element of interest.

Table 4-2 describes the available controls.

The following figures provide some examples of the different choices you can make with controls in the Performance Detail page:

• Figure 4-16 displays CountLoanOffers KPIs in a line chart over a period of days. KPIs are not broken down by dimension. Figure 4-17 displays the same data, along the ProviderName dimension

• Figure 4-18 displays the CountLoanOffers KPIs for the ProviderName dimension and both dimension values (Star Loan Provider and United Loan Provider) in a clustered bar chart over a period of days. Figure 4-19 displays the same data, in a stacked bar chart.

In addition to using the controls to view data along different dimensions, when hierarchical dimensions are defined, you can drill down in the dimension by clicking data points in the chart. A hierarchical dimension is one that uses ordered levels as a means of organizing data. For example, in a time dimension, the hierarchy organizes data at the year level (hierarchy 1), the month level (hierarchy 2), or the day level (hierarchy 3).

Figure 4-20 shows the data presented along the CarMake dimension – a hierarchy 1 dimension. Figure 4-21 shows the result when you click a bar representing the BMW make; all of the BMW models are presented.

Viewing and Understanding Cause/Effect Tables – Performing Root-Cause Analysis

A Cause/Effect table is one of a number of tables and charts that can be configured by the administrator to include in the Dashboard. A Cause/Effect table is useful for performing root-cause analysis of a KPI value.

There are two forms of the Cause/Effect table: a full form of the table, presented in a KPI detail page; and an abbreviated form, which is presented when a Cause/Effect table is included in a multiple pane view.

To access a full form Cause/Effect table:

1. In the Dashboard, select a multiple pane view from the Switch View box.

2. In the multiple pane view, click a data point in any chart or table. A Performance Detail page opens.

3. Click Cause/Effect.

4. In the Switch KPI box, select the KPI for which you want to view cause and effect details.

5. In the Analysis Time Grain box, select the unit of time (day, month, quarter, year) over which you want Cause/Effect details presented.

Figure 4-22 shows an example of the full form of a Cause/Effect table. (The abbreviated form excludes the Performance, Cause/Effect, and Compare columns.)

A Cause/Effect table shows the following, in relation to a selected KPI:

• Cause KPIs – Each of the KPIs that influence the selected KPI's value, the amount of influence each Cause KPI has on that selected KPI, and the KPI performance band value for the Cause KPI are presented. In Figure 4-22, for example, the SumLoanAPR value has an influence of 30.89% on the CountLoanOffers4United KPI value. The SumLoanAPR KPI is under performing for the day (as indicated by the Band column).

• Effect KPIs – Each of the KPIs on which the selected KPI has an influence, the amount of influence the selected KPI has on each of the Effect KPIs, and the KPI performance band value for the Effect KPI are presented. In Figure 4-22, for example, the SumLoanAPR KPI's performance has an influence of 31.67% on the AvgLoanAPR KPI's performance. The AvgLoanAPR KPI is a normal performer for the day (as indicated by the Band column)

In addition, in the full form of the table, the Performance, Cause/Effect, and Compare columns provide links to detail pages about each of the KPIs presented in the table, as follows:

• Performance

  Provides a link to the KPI Performance detail page for the associated Cause or Effect KPI. If the Band column indicates that the KPI performance is other than expected, you might click this link to see how that particular KPI has been performing, particularly if it has a strong influence on the selected KPI or is strongly affected by the selected KPI.

  See "Viewing and Understanding Performance Details" for more information.

• Cause/Effect

  Provides a link to the KPI Cause/Effect detail page for the associated Cause or Effect KPI. If the Band column indicates that the KPI performance is other than expected, you might click this link to see what is influencing this Cause or Effect KPI and what other KPIs this Cause or Effect KPI is affecting.

• Compare

  Provides a link to a page that presents the selected KPI and the Cause KPI in bar charts for easy visual comparison. Clicking the Compare All button displays all of the KPIs accessible to the user in bar charts.

  Figure 4-23 shows an example of the page that opens when you click a Compare icon. The page that is presented when you click the Compare All button is similar, but includes a chart for each of the Cause or Effect KPIs.

Viewing and Understanding KPI Context Documents

You can view a list of documents related to the KPI you select, assuming the administrator configured them, in the Information Documents page. Documents are listed in order of their weighted value. The weighted value indicates how likely the document will contain information of interest to you.

To access KPI context documents:

1. In the Dashboard, from the Switch View box, select a multiple pane view.

2. In the multiple pane view, click a data point in any chart or table. A Performance Detail page opens.

3. In the Switch KPI box, select the KPI for which you want to view KPI context documents. The Performance Detail page opens for the selected KPI.

4. Click Information. The Information Documents page opens and displays the list of information documents and their weighted values, as shown in Figure 4-24.

Viewing and Understanding KPI Predictions

The Dashboard Prediction Details page provides a prediction for a selected KPI value, based on related KPI values.

To view KPI predictions:

1. In the Dashboard, select a multiple pane view from the Switch View box.

2. In the multiple pane view, click a data point in any chart or table. A Performance Detail page opens.

3. Click Prediction.

4. In the Switch KPI box, select the KPI for which you want to view a prediction. The Prediction Detail page opens for the selected KPI.

5. In the Analysis Time Grain box, select the unit of time (day, month, quarter, year) over which you want Prediction details presented.

Figure 4-25 shows a sample Prediction Details page.

The Prediction Detail page identifies the selected KPI for which prediction details are being presented (in Figure 4-25, CountReq), and a table of KPIs that Oracle BPEL Process Analytics used in calculating the prediction. These KPIs are referred to as source KPIs.

The prediction details for the selected KPI, include the following information:

• The predicted value for the current time grain. In Figure 4-25, the time grain is the current day.

  Oracle BPEL Process Analytics arrives at the predicted value for the selected KPI by determining the sum of all of the predicting KPIs' predicted values, each weighted by its confidence value. Using the values in this example, the equation would be:

  (4715 x (9/(9+35))) + (4680 x (35/(9+35)))= 4687
  
  

• The target value, if a target value was specified. For the KPI in Figure 4-25, a target value of 290 was specified.

  The target value is set by the Oracle BPEL Process Analytics administrator.

• The confidence level in the predicted KPI value, if a target value was specified. For the KPI in Figure 4-25, the confidence level is 31%.

  The confidence level is the average of the source KPIs' confidence values.

• The performance band in which the predicted value falls.

Oracle BPEL Process Analytics determines the set of source KPIs to use to form the prediction based on how closely they are correlated with the KPI for which a prediction is being made. If the absolute value of the correlation of the KPI being predicted and another KPI is .8, then that KPI is included in the set of source KPIs.

The values for each of the source KPIs are determined, as follows:

• The predicted value for each source KPI is determined by simple linear regression of its previous values.

• The correlation factor for each source KPI is a number between -1 and 1, as determined by standard statistical equations. It indicates how a source KPI affects the selected KPI.

• The confidence level for each source KPI is its correlation factor squared.