Bringing Focus to Data Exploration with OEID, part 1 of 2

Oracle Endeca Information Discovery (OEID) offers two standout elements.  The first is support of “Unstructured Data”, I’ll speak to that in another posting.  The second is the use case around “Exploring Data”.  Our partners have enjoyed tremendous value from being able to navigate through large volumes of data using OEID or its previous release as just Endeca.  The tool allows analysts and executives alike to browse through all their data, and drill into areas by whatever catches their eye.

This is accomplished using what’s called the “Navigation State”, this maintains the context of where they’ve browsed to using “Breadcrumbs”.  This is also very similar to how many Business Intelligence tools allow you to cross drill along different dimensions.  You’re exploring your data, and that can be incredibly insightful for your business.


In reality however it can be a hard sell describing to some users that with only a few dozen mouse clicks they can drill down to a set of records that hold interest to them.  The free form act of “exploring” can be time consuming, and not every business user has that luxury of time.


OEID offers a couple great little features that address this challenge, the first I want to present are “Alerts”.  These are updated as part of the data store, and like all the charts and graphs and other OEID portlets automatically apply your current navigation state.  You might need a bit of technical help to create an alert, but they can flag specific conditions you want to monitor and highlight them for you to drill straight to directly.  You click once and all the relevant filters or breadcrumbs will be populated for you.

For example, let’s say you sell shoes and want to track which store(s) have failed to hit their sales targets.

Step1:  You create an “Alert Group” and define a SQL-type of query.  It’s not ANSI-SQL, but it can run calculations, group the results, apply case statements and where clauses, all the familiar standards for data analysts to ask questions by. 


Step 2:  Format the resulting record set values as percentages, currency, etc to apply business friendly formatting.  This is optional, just good style.


Step 3:  Lastly you write the actual alert message to display your result values or groups in whatever readable or information packed form you like.


This particular query isn’t complicated, but neither is the process to add this to your OEID portal.  The Alerts component will display in descending order ALL the records that would be returned to that query.  If your user has been exploring around the “nav state” the alerts will automatically filter them against the matching records.


When the user opts to click on one of the alerts all the groupings you’ve defined will be automatically update your navigation state and the targets records will be displayed.  You can even redirect your user to a specific page on your OEID portal where different portlets can display more detailed record values.

Effectively right off the first page you or your executive has been provided with a highlighted set of records that might need attention.  Of course you could still browse your filters to drill down to those same records, and that will remain a great way to explore and understand your corporate data.  Alerts is simply a very convenient feature if you’re short on time or not interested in the idea exploring millions of records.


The second item I wanted to highlight is the “Bookmarks” component.  As with all the OEID portlets this little gem can be dragged onto any page and automatically integrates with the page through the navigation state.  In fact it is uniquely tied with the navigation state and expands on how a user might have explored the dimensions (i.e. breadcrumbs) to arrive at a certain point.

Any OEID user can create a bookmark, just as easily as they bookmark a website.  You only define a name and description and it will capture the filters or breadcrumbs you have currently selected.  Nothing more complicated than that.  In terms of functional use however this is a wonderful convenience.

Firstly you can avoid remembering which filters you were applying to arrive at a certain data set.  Your navigation state has been saved and with one click on the bookmark will automatically be reapplied.  Even the page you were on will be saved.  Speaking for myself the less I have to remember the better!

Secondly you can monitor a data set.  With the example I used for the Alert above you could create a bookmark to check back in and see how that store was doing.  In a later month the alert conditions may no longer be met, but might still want to see exactly how things have shifted.

Thirdly you can share your bookmarks with.  They’ll need access to the OEID portal, but if you do the analysis to find a store that needs attention, just create that bookmark and email it to your boss or colleagues so they can zero directly in on a record.   No time wasted, no repetition of steps, no fuss!

There are two overriding value adds that “Alerts” and “Bookmarks” add.  Firstly they work seamlessly together.  This is a simple and reliable pattern you’ll find repeats throughout OEID.  The integration of the components is part and parcel of their design and their implementation.  The can exist on their own, but they also complement each other very nicely.

The second is they are exposed.  Writing a query for an Alert isn’t going to be an appealing task for your non-technical users, but your power users, data analysts and Excel guru’s will figure it out.  There is no secret IT magic that has to happen behind the scenes.  These features are fully functional out of the box and right on the interface.

Alerts and Bookmarks are nice little features that you have to start using.  They don’t take away your ability to explore your data.  They just let you bring a little focus to it.

This entry was posted in Endeca and tagged , , , , , , , , , , . Bookmark the permalink.