I've got another gig on Fridays with the goal of building an analytic app. I had been working things out on the side to load data into mysql. I started building canned reports with Ruby on Rails. There were some things l couldn't do easily with Ruby and I ended up looking for a solution that already had thought about slice and dice, drill down capability reports. Maybe 5 years ago, your options would be one of the expensive commercial packages - Cognos, Business Objects, Crystal Reports, Brio, Sagent. Your database 5 years ago would have had to be Oracle, Sybase, Sql Server or DB2. Luckily today, I can choose from a number of open source DBs like Hypersonic, Progress, and MySQL. Progress and MySQL have been around for a while. Progress I believe is rooted from the old Ingres database which I used when I was in college back in the early 90s.
For the reporting and ETL end, I started looking at Pentaho. Some of my old Sagent cohorts found a way to this company to ties together several open source ETL and reporting projects into a BI Suite. Most of the products are Java based and are tied together through the Eclipse IDE (integrated development environment). To hear Eclipse and Java can be intimidating and a barrier to entry to some. I'm ashamed to say that at first, I shied away from the technology because I didn't want to get into something that I needed to be a Java person to do. I've been writing little Java programs hear or there and have been a member of the Philly Java Users Group for years. I've never written Java code for a job. I've been using Sagent, some perl scripting, some stored procedure writing, and a lot of vi with regular expressions to shove data into databases. It turns out that though everything is built on Java, the interface (which is constantly evolving) is more of a graphical process diagram where you fill in the information to make things work. The diagram includes inputs, actions, and outputs. For inputs, you can choose from flat files, SQL, XSQL, or XML files. For outputs, you can write to a database, flat file, or XML file, a formatted report or OLAP report via HTTP. In between, there are a number of different options for transforming the data.
I thank my patron for buying me the time to spend looking at this. I was resistant at first but I see the beginning of a beautiful and potentially fruitful relationship between myself and open source BI. My goals for the rest of the year are to help my patron build his product. If the product takes off, I may specialize in the product. If it doesn't or even if it does, I may specialize in the process of offering open source BI to folks in the Philly area and the surrounding Tri-State.
It's not out of the question as well for me to join the Pentaho company. I have my contacts into the organization and my parents live in Florida where the company is based. That's in the back of my mind but for the past 10 years now, I've set roots for myself and my children in the Philadelphia area.


