Read our overview of the Keynotes
Read our overview of the talks in the Technical room
Devan kicked off the talks in the business room by sharing his experiences in building a health data in Mozambique.
Part of the scope of the project was the geographical allocation of e.g. nurses over Mozambique, which poses some specific challenges in a developing country like Mozambique.
The reasons Pentaho was chosen are mainly because the need for
The project team faced some specific challenges in the implementation, e.g. people are heavy paper users, therefore all reports (including a 220 page report) need to be printable.
A couple of issues the team faced during the implementation:
Stefan Mueller started his talk with an introduction of IT-Novum as an open source integrator and SAP partner.
To support their customers in this combination of open source and SAP, IT-Novum created 4 plugins to use SAP data in PDI:
All steps support metadata injection. Stefan wasn't clear on whether the steps will end up in the PDI marketplace, but ensured the audience the steps can be supported commercially to the user community.
Kamil, who is a researcher at the Technical University of Liberec (Czech Republic) introduced their hydrogreological information system, available at dataearth.cz.
To design ground water models, lots of data need to be integrated and analysed. This data comes in both structured and unstructured forms, but always with a strong geographical angle.
The data is loaded into star/snowflake schemas. The project uses the entire Pentaho stack, but because of the nature of the data, are heavy users of the PDI Gis plugins.
These schemas are then used to perform analytics on e.g. water iron levels, water quality, geographical mappings etc.
With the NOVA project, Essent Belgium is in the process of migrating the entire software stack from on-premise Microsoft based applications to open source based software in the cloud (AWS).
I talked about the agile project structure, the as-was and as-is architectures. In the new architecture, DMS and PDI are used to load data from over 10 source systems over landing and logical layers to a Redshift data warehouse.
From the logical layer, data is split of (Nifi) to an analytics layer, where a team of data scientists build models using R and Spark.