On October, 24th, just weeks after PCM13, the first Pentaho Benelux User Group meeting took place at the Antwerp Train station, organized by Matt Casters and Bart Maertens, and sponsored by know.bi and Pentaho. About 60 people gathered there for presentations, and to meet each other over beer and pizza.
Pedro opened the presentation slot with a talk about Pentaho and the company's relation with its community.
After that introduction, he quickly jumped to discuss the newest developments in CTools. CTools are Pentaho's set of power tools that enable developers to create amazing dashboards.
The latest additions to the tools allow developers to create operational analytics, so a dashboard user can place an order right from a dashboard.
The biggest newcomer in the CTools family is Sparkl. Sparkl, apart from breaking the C** naming convention, is an application builder for the Pentaho platform. Using Kettle to disclose data sources and dashboard technology to create user interfaces, Sparkl allows developers to create applications in a matter of hours. This and makes Pentaho the first BI platform with an app store, and could be a real game changer.
After Pedro, Michiel Brunt and Maik Groenewegen took the floor to talk about how they use Kettle.
After an introduction about Inergy, Michiel explained the Inergy architecture (Kettle loads data in Netezza, Microstrategy frontend). They stressed the need for automation, to design and develop for failure, with a strong focus on job restartability. To manage their development process, Inergy uses an agile methodology. Keeping track of data history (handling Slowly Changing Dimensions) is a very important aspect of their implementation, but is easily dealt with using Kettle's 'Dimension Lookup/Update' step.
Michiel and Maik discussed their tool selection process (which Kettle obviously won), and talked about their installation, deployment and versioning process. A number of hints on best practices concluded this presentation of what seems to be an excellent implementation.
Comways' Jorge Alamanac started his presentation about the Euroconsumers BI project with a short introduction about Comways, a service provider for contact centers.
For this project, Euroconsumers needed to load data from 5 geographically distributed Altitude servers to a central data warehouse. The data needed serves 16 KPIs, is loaded daily, and is used by 50 users in different time zones all over the globe. The entire project runs on the Pentaho EE stack, with PDI to do the heavy data lifting, Mondrian/Analyzer for self service reporting and a number of predeveloped reports (some of which are bursted). Active Directory is used for authentication and authorization. Jorge then dived into a couple of challenges they faced during the project, mainly related to PDI and Mondrian performance, and enforcement of user security. Finally, Jorge mentions that the solutions built are used not only as a pure BI solution, but also for forecasting and workforce planning.
After a short break from the hot room, Kettle founder and community member since day 1 Matt Casters took the floor to talk about the new features in Kettle and Pentaho 5.0.
Matt doesn't do powerpoint presentations, but actually showed the new functionality in PDI.
Just to name a few:
As with every new release, there are new transformation steps and job entries, and even more can be fetched from Kettle's marketplace.
Nele Dekien, prominent member of the Cipal Pentaho team, came to talk about how Cipal uses Kettle to gather data from customer servers. Cipal, a provider of IT solutions for Belgian governmental organizations, has a number of applications installed onsite at customer infrastructure. Because of these local installations, their ETL processes can't access the data directly. To solve this, Cipal developed AthenaExtract, a Kettle based solution. Kettle is installed on the local customer server, containing just 1 job that fetches the 'real', application and version specific ETL code from the Cipal FTP server. These ETL processes are then executed to extract data from different application databases and upload it to Cipal's FTP server, where it is loaded into the data warehouse.
Aly Van Zalk, long time community member and scrum master for the St Antonius Hospital in Utrecht took the floor in the fist part of a duo presentation.
Aly talked about Pentaho's community, which she described as the company's secret weapon, and mentioned some of the contributions the hospital sponsored. After that, Aly discussed the hospital's KPI management framework, a graphical user interface to control dashboard KPI's. These dashboard were developed using WebDetails' CTools, and show some stunning visualizations. Aly then discussed some of the latest developments in the hospital and then passed the floor to Edwin Weber. Edwin introduced his Data Vault management framework, developed in Kettle and confugured through Excel spreadsheets. Since the hospital is going through a round of savings, all external contracts are killed of, so Edwin is forced to document his framework and teach the internal staff how to use it. Edwin told the crowd about the contacts he had with other BI vendors to port his framework, and mentioned the download site at sourceforge.
After a quick round of Q&A, the fun part of the afternoon started: beer, pizza and informal talks on the wonderful balcony of the train station, with a glorious view on arriving and leaving trains.