In choosing a tool like OBIEE companies are looking for more than a reporting solution. They want users to gain insight into their data and find anomalies and trends hidden in the transactional data. However, the tools available so far in OBIEE have really divided users into those that can learn and excel in Answers (Analysis), and those that are confined to dashboard consumption of data. In Fact Gartner finds that:
At most of my clients many users are tied to the availability of report developers or power users to get the data they need. Many times it is a new attribute or a different aggregation or sort that is required. With the new dashboard features in OBIEE 11.1.1.6 a whole new group of users can have the power of truly interactive reporting by planning for and incorporating content into dashboard for that purpose. This article will outline the methodologies that can be incorporated in dashboard design and the basic training it would take to get that next set of users comfortable with grabbing their own data thus enabling them to accomplish their jobs easier, and freeing up IT resources for more complicated tasks.
What Actions Are Available
In the properties section of each analysis is a tab that allows developers to set the availability of various report manipulation features. The list below and indicates what the features can be allowed for dashboard users to manipulate a single report.
Drill
Drilling has been with OBIEE since the beginning. For the end user it allows the ability to get to details quickly and to zoom to the results of the same report on fine grains of data. This has been setup up as the primary interaction on the column properties for the year column below. The change in 11.1.1.6 is the ability to drill by right clicking to the interaction menu.
Moving Columns
Let’s look at using a table view in a dashboard. For this example I created an analysis with a couple of hierarchies and many hidden columns. The use case is to provide a single report that can answer many questions about that subject area and is customizable by the end user.
Here is the base line report saved on a dashboard. Let’s explore a few use case scenarios an end user may want handle without going to the Analysis tool.
Prompting and Sectioning:
This is a straight forward example that has been available since 11G came out but let’s walk through the use case. A user wants to be able to prompt on any dimension column in a report. Let’s say Brand for our example, which is a matter of dragging the desired column to the prompt drop space on the report or selecting move column à To Prompts in the right click menu. You will notice in this snapshot that the right click menu got much bigger and we will go through the list one at a time.
Moving this to a prompt condenses the report and allows the user to work with one brand at a time
Or with a section:
All these reconfigurations of data are truly useful to end users that need different views of data but are not comfortable utilizing the Analysis tool. The ability to turn off the feature is equally important for some reports. Dashboards created with many reports formatted a certain way or for public consumption may make sense to disallow this feature.
Sorting Columns
Sorting is another feature that came with 11g originally but now can be controlled in the report properties. Users can change the sort of a column on the fly showing numbers in a way that makes the most sense to the individual user.
Add and Remove Values
The add and remove values feature is set up in 2 ways. At the attribute level a user can remove or keep only the selected value or values. This is useful if a user only cares about certain regions or just does not care about certain values and want to clean up their view of data.
The more powerful aspect of add and remove feature is at the column level. Here a user can enable some complex comparisons and analysis. For each dimesnion value the users can select to keep or remove based on any measures top X or bottom X. So for example say this users is only interested in years with the top 3 based on total damages. These kind of dashbaord based analysis can really enable a busienss end user with simple but powerful tools. As long as the measure is part of the criteria the user can manipulate the results on it.
The user can also filter the report based on comparing columns. Say the end user want to only see years where the number of fatalities is more than the number of hospitalizations, or a measure based on a fixed value. The possibilities of filtering and calling out important data are really endless with this feature. Like all these customization interactions, a user can save important ones as their dashboard customizations.
Groups
To further customize my view, I can remove the details behind my group by removing the departments that start with T rows. Notice the grand totals stay correct no matter how many groups have been created.
Calculated Items
Calculated Items allows for the addition of rows to the report with customized aggregations. For example, we can add a row for the average office. By control clicking all the office members and then right clicking and selecting add calculated item.
The results show a new row with the average for the measure in the table.
Subtotals and Running Sums
The ability to add totals, subtotals, and running sums to any report adds another layer of customization for the dashboard ad-hoc user. In this example a report starts with not totals or subtotals and by right clicking the end user can customize a report that shows both at any level.
First Grand totals, like in the pivot table editor the user can choose to add them after or before the columns or rows, giving are dashboard only user the abilities to customize the report like a report developer.
Users can also include subtotals when 2 or more dimensions are on an axis.
Finally the user can pivot the measures to be absolute or running sums.
Include and Exclude Columns
To truly enable ad-hoc analytics as a dashboard only user, the report builders should consider adding hidden columns that can be included or excluded depending on the end users preferences. Let’s look at an example with Sample App using a report based on Time, Product, Office, and 3 measures.
Traditionally to save real estate we may design the report with column selectors or view selectors, maybe variable prompts or other methods for controlling what attributes are available. But now we can design a report with excess columns and give the presentation choices to the end user directly. First let’s look at the measures; we can include others by right clicking in the measure section of the pivot table.
he same can be done for dimension columns in excluding or including them, then saving the customization in the dashboard.
With the powerful right click options available to dashboard end users, report developers can rethink the content of reports to truly expand the possibilities for end user ad-hoc report adoption without having to train on the analysis tool. This opens the door to a new class of user who is curious about the data but not ready to commit to ad-hoc development.
