Less than two weeks after upgrading the SpagoBI software that powers the Visual-Baseball Analysis Lab, I’ve decided to move in a different direction. After spending many hours over the holiday weekend examining all the options, I’m electing to move forward using d3 and nvd3 as the primary tools for the future version of the lab.
Here is some rationale behind my decision:
- SpagoBI, being a java-based platform, has a huge footprint, which means that I need to maintain a Tomcat web server with plenty of memory and storage space. Given the limited usage of the lab over the last two years, this no longer seems like a good tradeoff.
- It takes too long for the app to launch after logging in. I could spend more time addressing this, but again this is not time well spent. I would rather have an application (or applications) that are fast and easy for site visitors.
- Java-based platforms, including SpagoBI, Jasperserver, and Pentaho, all lean toward production-oriented reporting. This makes sense, given their corporate audiences, but is no longer the best option for what I hope to accomplish with the VBP site. I need a less rigid model with greater growth potential.
- Most of my recent efforts have been created using d3 and nvd3, so it makes sense to leverage these tools even more, and to spend a higher percentage of my limited time using the most effective tools.
I will miss certain elements within Spago, and in the general BI model, such as OLAP cubes and parameterized reports. Perhaps these will reappear in some form in the future. On the flip side, I certainly won’t miss stack errors, re-booting Tomcat when the app crashes, and a few other annoyances that seem to be standard fare with Java. There are still some worthy Java apps, including Spago, but the time has come to move forward. More to come.by