Analytics as a Service

Company overview

CatMan Solution [CatMan] was established in 2002 (A-Ware Software Development) and has become a leading specialist in the Nordics on POS data for the FMCG sector. CatMan specializes in the field of Business Analytics for national and international FMCG customers. Our analytical suite created exclusively for CPG companies is available in subscription format.

Implementation summary

We have managed well over 200 different data sources within our service platform. The latest step has been to migrate some of the workloads on to a Fabric based platform, away from a SQL Database Elastic pool solution. The solution is not only based on Microsoft Fabric, but a number of Azure services as well. We have choosen Azure Logic Apps as a thin client to bring in data from protocols like mail and s-ftp. These land attachements into a storage account, where we register the event using Azure Event Grid. These blob created events will trigger of a registration of each blob, and this again fires of a new set of events to see if we are to process the file at all (Firewall), is the file in a compressed format (UnZip) and some files are even in a corrupt state, so we can rectify that. Once this has been completed as classification (Classifier) of each files is undertaken, and the outcome of that is a reference to a specific Notebook in a specific workspace with a specific set of parameters that gets executed (Executor). Every classification has a sort of a flight plan, dictating how the data should travel through the medallion architecture layers.

With the initial set of customers loaded, we are looking at a silver lakehouse, hosting the combined set of data, that has a size of about 50 GB. This is broken up into customer focused gold lakehouses, containing only their own data.

Business impact

• Greater speed to insights (V-Order, Power BI)
• Seamlessly integrate data in “one location” (OneLake)
• Out-of-the-Box support for a great variety of sources.

Key deployment drivers

• Faster load times (end to end)
• Broader adoption of technology in organization
• New opportunities w/ Fabric
• Support for larger amounts of data

Deployment team

• Simon Fischerman, Developer, Architect
• Thomas Tang, Developer
• Anne Jylov Grosch, Developer
• Jens Vestergaard, Principal Architect, Developer

Architecture diagram

Expansion plans

• Expand to support all of the 80 POS source formats we currently host on the SQL platform today
• Monitoring, Automation and Configurability
• Support for multi Fabric Tenants (B2B monitoring and operations)