OpenI plug-in adds following key features to Pentaho community edition:

Login Behavior

When users login, OpenI shows a default dashboard. It also enables multi-tenancy where multiple projects can be hosted on the same instance of the Pentaho server. So, depending on which project a user belongs to (managed in user configuration), they will see their corresponding dashboard.

OpenI User Interface to Build and View OLAP Reports

  • Simple, clean layout built on top of jPivot
  • Replaces jPivot's default buttons and icons with cleaner selection check boxes and icons
  • Navigation on the left - provides drag and drop to move dimensions and measure from rows to columns to filter section
  • Can collapse/expand left navigation bar to increase screen real estate for report
  • Table and chart view on 2 separate tabs
    • Prevents over-crowding of X-axis labels
    • Charts fit to available screen real estate
  • Shows filter criteria selection on both chart and table view
  • Allows selecting multiple values for Filter dimensions
  • Cleaned up pop up window interface for report configuration properties (Chart, MDX, Print, and Sort)

Explore OLAP Cube Data

Almost all OLAP reporting interfaces assume that the users know which measures they want to view and which dimensions they want to see on columns and rows. As such, as a starting point to viewing any OLAP data, these tools always ask the users to select dimensions to put into rows and columns and select the measures. However, most of the time, the users just want to browse what's there. The "Explore Data" feature in OpenI enables just that:

Users select a Cube to browse, and can optionally also specify a particular set of measures (metrics) they are interested in exploring (handy in cases where a cube has many different measure values). Then the users click on Explore - that's it!

OpenI will show a dashboard like view where the selected measures are charted with respect to every single dimension in the cube so that the users can "eyeball" in one shot how the particular measure(s) varies by each different dimension values. From here, the users can decide which particular dimension view they want to drill into further. So, they click on the particular chart, which will then take the user to the detailed OLAP report view, with the selected measures and dimension already placed on appropriate rows and columns -- so that the user can then select how they want to slice and dice the data next.

So -- this feature provides more of a top-down approach to building OLAP reports -- first, provide a comprehensive view, and let the user visually select which detail they are interested in, and use that as the basis of a new report definition.

Custom Drillthrough

Typical behavior of OLAP drillthrough is let user download a data set of all the identifier values (like customer_id, or order_id for example) belonging to a particular cell in the OLAP report. In practice, this is not that useful, because most of the times the users need more data than just the identifiers. Also, a lot of the times, the user does NOT want to download all the data to their personal machine, because their end goals is to actually submit that data to an external system such as another 3rd party application or maybe an FTP site.

Drillthrough in OpenI is called "Data Report" - clicking on it will show the green arrows next to each cell value in the Table view. When users click on a green arrow, a pop up box comes up asking whether they want to submit the resulting data set to a web service, or download -- so, this avoids always having to download the data to a local machine.

Furthermore, in OpenI configuration, administrator can write a custom SQL query for drillthrough. This way, the drillthrough result can be an entire result set of the custom SQL query. For example, instead of just returning customer id's, you can return all the extended customer profile data such as demographic data, contact information, etc. and you can also add additional data points such as the orders the customer has submitted, etc. Of course, you have to take into consideration on how all this may bloat the result data set, but the key point here is that you have full control over the format of the data set returned by drillthrough, AND the user has full control over whether they want to download the results to local machine or submit it to an external application.