WIT has been using Argos since 2007, but mainly to just get spreadsheets of data from Banner. This past year, we started a new initiative to really help visualize the data and make it easier to get what you really want out of Banner.
The impetus for this was the amount of manual manipulation of spreadsheets we heard staff were doing. In some cases, it could take up to 2 weeks to gather a list of GPAs for a particular major for particular courses. We set out to create our own design language for datablocks going forward – previously, we had just used a simple gray background with one filter slapped on. Now, we have our own branding and template we use for each new report.
In conjunction with the new design, we also utilized Oracle to help build the custom tables the datablocks use for queries. We were finding 2 issues with our old method: scalability and speed. For scalability, our new method sometimes requires up to 12 filters to make it useful for everyone on campus. Without writing the SQL in an Oracle package and refreshing the data every night, it would be impossible to implement the smart filtering system on all of our reports. Even with that, the filtering grows more complex every day.
The other issue was speed. By going directly against the baseline tables (or even ODS views), the sheer volume of data being queried was grinding our reporting efforts to a halt. By combining all of the data we want into one static table, reports that may have taken 60 seconds to run can do the job in under 10 seconds.
The main focus of the presentation is how we went from using Argos in a stale manner to making it a user-friendly, indispensable tool across campus (and how Oracle supports the back-end efforts with that). However, it would also include some workarounds and limitations we have found (such as charting issues with more than one sequence and having to explicitly create a smart filtering system, which makes changes down the line much more complicated).
Benjamin Assa, Data Architect